` element decides your Google click bait. Aim for a specific, 30–60 character title on every page.\n- **Meta descriptions missing or duplicated** — The one-line snippet Google shows below the title. Write a unique description per page.\n- **Page headings missing or overused** — One `<h1>` per page, then `<h2>` / `<h3>` for structure. Missing `<h1>` is a big deal; two `<h1>`s is a smaller one.\n- **Canonical link missing or broken** — A `<link rel=\"canonical\">` tells Google which URL is the \"real\" one. Missing or pointing off-site is bad.\n- **Structured data missing or broken** — Machine-readable tags (JSON-LD) that help Google understand articles, products, FAQs.\n- **Broken internal links** — Links from one of your pages to another that return an error. Fix or remove them.\n- **Thin content** — Pages with fewer than ~200 words look empty to Google. Expand the ones that should rank.\n- **No FAQ signals found** — No question-and-answer sections detected. FAQs help both users and AI assistants.\n- **llms.txt file missing (advisory)** — An emerging convention: a `/llms.txt` file that hints AI crawlers about your best content. Not required — advisory only.\n- **HTTPS not enforced** — Your site should redirect http → https and canonicalise the https version.\n- **Poor real-visitor performance** — Real visitors experienced a slow page. Data comes from Google’s Chrome UX Report — sites with little traffic may have none of it yet.\n- **Low lab performance estimate** — Our synthetic Lighthouse run scored low. A single run varies ~10 points; treat as a hint, not a verdict.\n- **Not mobile-friendly** — Phones do not render the page well. Check tap-target sizes and the viewport meta tag.\n- **Not indexed by Google** — Google looked at the page but did not add it to search. Reconnect Google Search Console if this rule cannot fetch data.\n- **Rich results issues** — Structured data was recognised, but Google flagged errors that block enhanced search results.\n- **Indexed with warnings** — The page is indexed but Google flagged a canonicalization or duplicate-content signal.\n\n[Back to the docs index](./index.en.md)"},{"slug":"troubleshooting","locale":"en","title":"Troubleshooting","description":"Common problems and how to fix them.","section":"start","order":3,"searchText":"# Troubleshooting\n\n## Audit failed or unavailable\nAn audit fails when we cannot reach your site or the crawl provider timed out. Common causes:\n- Your site is behind a Cloudflare \"under attack\" rule blocking crawlers.\n- Robots.txt blocks user-agent `RankMeFastBot` or \"`*`\".\n- Origin returned 5xx for our first requests.\n\nTry again in a few minutes. If it keeps failing, adjust firewall/robots.txt to allow our bot.\n\n## Rank check unavailable\nA single keyword can come back unavailable if the SERP provider is rate-limited or Google served a captcha. It is temporary — the next scheduled run retries. Nothing to do.\n\n## \"Reconnect needed\" on Search Console\nGoogle invalidated our token. Open Settings → Google Search Console → **Reconnect**. This can happen if you changed your Google password or revoked the app.\n\n## Cap reached\nSee [Plans, limits & credits](./plans-limits-credits.en.md). Options are: upgrade tier, buy a small credit pack, or wait until the next billing period.\n\n## Cap reached on a research surface (402)\nThe response names the exact unit that ran out — for example `keyword_lookups` or `audience_research_runs`. Wait for the monthly reset, buy a credit pack where one exists, or upgrade; Audience Research runs have no pack.\n\n## Partial results on a research surface\nSome engines or sources answered, others did not. The result ships with what arrived and is labeled **partial results** — the missing part is named, never zero-filled.\n\n## Unsupported for this engine / market\nThe check cannot run for that engine or market. The unsupported part is excluded from the result — it is never displayed as a zero.\n\n## Unavailable (not zero)\nNo data came back for the window. `unavailable ≠ 0`: treat it as unknown, not as a drop to zero. You will see this on share-of-voice and Search Console generative-AI cards.\n\n## \"Reconnect needed\" on analytics or generative-AI cards\nSame cause as the Search Console reconnect above — Google invalidated the token. Reconnect from the site's Google workspace; the cards recover on the next read.\n\n## Verification email missing\nCheck spam. The link expires after 24 hours. If it truly did not arrive:\n- Confirm you typed the address correctly (typos are the #1 cause).\n- Try Google sign-in instead if your address is Gmail.\n- Contact support if the address is a corporate one — the mail may be filtered upstream.\n\n## AI summary shows \"shortened\"\nLong audits are truncated before we send them to the model — the summary is still valid, just from a shortened view. Click **Regenerate** if you want another take.\n\n## Report screen loops loading\nHard-refresh the page (⌘/Ctrl + Shift + R). Clear the local cache if that does not help. If the issue persists, it usually points at a mismatch between an in-flight audit and the report snapshot — support can force-refresh from their side.\n\n[Back to the docs index](./index.en.md)"},{"slug":"changelog","locale":"en","title":"Changelog","description":"Plain-language notes about product changes that affect how you use RankMeFast.","section":"start","order":4,"searchText":"# Changelog\n\n## 29 July 2026 — New intelligence tools, no price change for existing subscriptions\n\nWe added [Link Intelligence](./link-intelligence.en.md), [Traffic Insights](./traffic-insights.en.md), [Keyword Trends](./keyword-trends.en.md), [Review Intelligence](./review-intelligence.en.md), [Brand Radar](./brand-radar.en.md), and the public [SERP Sensor](./serp-sensor.en.md).\n\nIf you already subscribe, your current RankMeFast subscription price is unchanged. The new monthly allowances, optional Brand Radar add-on, and one-time packs apply as described on [Pricing](./pricing.en.md). Add-on and pack prices apply only if you choose to buy them.\n\nWe are not moving you to another plan, and you do not need to take action. Your existing stored reports and history remain available under the rules of your current plan.\n\n[Back to the docs index](./index.en.md)"},{"slug":"page-speed","locale":"en","title":"Page speed","description":"Core Web Vitals in plain words — real visitors vs a lab estimate.","section":"audits","order":1,"searchText":"# Page speed\n\nGoogle grades every page on **Core Web Vitals** — three signals about how it feels to load and interact with your page. There are two flavours of measurement and RankMeFast shows both:\n\n## Real-visitor data (field data)\nGoogle’s Chrome UX Report aggregates real visits from Chrome users. When your site has enough traffic, this is the number that matters — it is what Google will judge you on. The three signals:\n- **LCP (Largest Contentful Paint)** — how quickly the biggest visible element appears. Aim under 2.5 seconds.\n- **INP (Interaction to Next Paint)** — how snappy your page feels when you click or tap. Aim under 200 ms.\n- **CLS (Cumulative Layout Shift)** — how much things jump around during load. Aim under 0.1.\n\nIf your site is new or has low traffic, field data may not exist yet — the row shows \"not enough visitor data yet\". That is normal.\n\n## Lab estimate (Lighthouse)\nWhen there is no field data, RankMeFast still runs a synthetic Lighthouse test in a controlled environment. Treat this as a **hint**: a single Lighthouse run varies by ~10 points from one attempt to the next. Do not chase small changes — look at the trend across audits.\n\n## Mobile\nMobile is measured separately because phones render pages very differently. If your page is not mobile-friendly (missing viewport meta tag, tap targets too small), it lands in **Fix now**.\n\nTip: page-speed problems are usually image-shaped. Compress big images, serve modern formats (`webp` / `avif`), and lazy-load anything below the fold.\n\n[Back to the docs index](./index.en.md)"},{"slug":"ai-summary","locale":"en","title":"AI summary","description":"What it does, what we send, and why it never changes the audit.","section":"audits","order":2,"searchText":"# AI summary\n\nThe AI summary sits above the tabs on the report screen. It reads your audit and writes a short paragraph in your language telling you which fixes to do first.\n\n## What it does\nThe summary is a plain-language executive brief: which findings are most important, why they matter, and roughly which one to start with. It does not add new findings — it re-tells the ones the audit already produced.\n\n## What we send\nOnly two things are sent to the AI model:\n1. The text of your audit findings — the same rule titles, plain-word explanations, and fix strings you can see on the screen.\n2. Your affected URLs, so it can reference them.\n\nWe do **not** send Search Console data, screenshots, your billing info, or anything about other sites on your account.\n\n## Why it never affects the audit itself\nThe audit is deterministic — same input, same output. The AI summary is a separate step that runs *after* the audit finishes, using the finished snapshot as read-only input. Turning the summary off has no effect on any rule’s bucket.\n\n## Regenerating\nClick **Regenerate** on the summary card to run it again. Each regeneration uses one summary credit (see the AI summaries cap in [Plans, limits & credits](./plans-limits-credits.en.md)). If you hit the cap the card shows a paywall notice; unpaid retries do not spend credit.\n\n## Choosing a model\nThe summary uses a small, low-cost Claude model by default. Some findings — very long lists — are shortened before sending. If shortening happens, the card shows a \"shortened\" note.\n\n[Back to the docs index](./index.en.md)"},{"slug":"content-intelligence","locale":"en","title":"Content Intelligence","description":"Analyze a page you own, score it, and get a plain-language brief with citations.","section":"audits","order":3,"searchText":"# Content Intelligence\n\nContent Intelligence looks at a page you own and returns a deterministic\nreadiness score, the evidence behind it, a plain-language brief, and — if\nyou opt in — an AI-generated draft. One analysis draws one\n`content_analyses` credit from your monthly allowance or your credit pack.\n\n## What it does\n\n- Fetches the page through the Firecrawl provider you own the crawl for.\n- Extracts headings, links, structured data, and a small set of on-page\n signals. No third-party HTML is stored raw.\n- Scores the page against the same rule engine that powers the audit report,\n so the numbers stay comparable.\n- Returns evidence for every finding: the rule id, where it fired, the\n citation excerpt, and a confidence label.\n- If you opt in, calls the AI provider order from your operator config to\n produce a brief and (optionally) a draft. The provider selection is\n server-side; you never see raw provider payloads.\n\n## What it does NOT do\n\n- It never publishes anything. Drafts are yours to accept, edit, or dismiss.\n- It never crawls arbitrary URLs. The URL must belong to a site you own or\n a competitor you have added on the Agency tier.\n- It never returns the raw provider payload, prompt, or completion. The\n public API response is the same normalized shape the app renders.\n\n## Starting an analysis\n\nYou can start an analysis from:\n\n- **Sites → any site → Content** — pick a page URL from your inventory.\n- **Report → Fix now / Watch** — the recommendation row opens the workflow\n pre-filled.\n- **Keyword research → tracked keyword** — analyzes the URL that ranks for\n the keyword right now.\n- **Google Search Console → any query** — analyzes the URL Google surfaces\n for that query.\n- **Competitors → Agency competitor page** — Agency tier only.\n\nEach entry point deep-links into the same workflow. Every one of them\nreserves capacity before enqueueing (see\n[Plans, limits & credits](./plans-limits-credits.en.md)); the flow refuses\nloudly if you are over your cap.\n\n## What one credit includes\n\nOne `content_analyses` credit covers:\n\n- one page fetch through Firecrawl,\n- the deterministic score and evidence,\n- the plain-language brief,\n- and — when opt-in AI is enabled — one draft.\n\nRegeneration draws another credit. Retrying the same URL within the\nseven-day snapshot window reuses the cached page but still costs one\ncredit for the new score and brief.\n\n## Score, evidence, confidence, citations, brief, draft\n\n- **Score.** Deterministic 0–100. Same rule engine as the audit report.\n- **Evidence.** Each finding names the rule, points at the exact HTML\n location it fired on, and shows a short excerpt.\n- **Confidence.** `high` / `medium` / `low`. Confidence is a function of\n page structure quality — it is not vendor spin.\n- **Citations.** Every AI claim cites the URL and excerpt it came from.\n Uncited claims are not shown.\n- **Brief.** A short, plain-language plan you can share with a writer.\n- **Draft.** Opt-in. The AI provider order in operator config decides who\n writes it. You accept, edit, or dismiss it — it is never auto-published.\n\n## Opt-in AI processing and retention\n\nThe AI brief and draft are opt-in. When you opt in, we send only the\nexcerpts needed to produce the brief. Excerpts are retained for seven days\nso we can regenerate on request without a second fetch, then purged. Your\nexport and account-deletion controls in\n[Account security](./settings-security.en.md) still apply and complete\ninside the documented grace window.\n\n## Partial and refunded results\n\nSometimes Firecrawl or the AI provider degrades. The analysis surfaces\nwhat completed successfully with a `partial` flag; you can regenerate the\nmissing sections at no extra credit. If the run failed hard before any\nuseful signal, the credit is refunded automatically and the analysis is\nmarked `refunded`.\n\n## Recommendations and 28-day correlations\n\nThe brief lists recommendations. You can:\n\n- **Accept** — moves the recommendation into your outcome loop.\n- **Dismiss** — records the dismissal reason for the correlation window.\n- **Apply** — marks it applied; the outcome loop measures the 28-day\n before/after correlation on rank, clicks, and impressions.\n\nThe correlation is descriptive, not causal — it is the same window we use\nin the audit report, and the number tells you whether the change moved.\n\n## Pro inventory and cannibalization\n\nOn Pro and Agency, the workspace opens a page inventory that groups URLs\nby topic and flags cannibalization pairs (two of your own URLs competing\non the same intent). Inventory scans do not draw analysis credits — they\nrun on the shared vendor cache.\n\n## Agency competitor portfolio and monitoring\n\nAgency adds a competitor portfolio: you pick competitor URLs, run analyses\nagainst them, and — with Content Monitoring on — get notified when a\ntracked competitor page changes materially. Monitoring uses the Firecrawl\nchange signature; a dismissed signature does not re-fire.\n\n## Managing allowances and the 20-analysis pack\n\nYour monthly `content_analyses` allowance depends on your tier — see\n[Plans, limits & credits](./plans-limits-credits.en.md). If you run out\nmid-month, the 20-analysis credit pack adds twenty single-use credits that\nnever expire and stack on top of the monthly quota.\n\n## Troubleshooting\n\n- **Access.** The URL must belong to a site you own (or an Agency\n competitor). Otherwise the flow returns \"not found\" — never \"forbidden\"\n — so it does not leak whether the URL exists.\n- **robots.txt.** If the page blocks crawlers, the analysis returns\n `partial` and cites the block; we never bypass robots.\n- **Stale GSC.** If Google has not yet indexed a URL, the analysis still\n runs on the fetched HTML but does not correlate against zero GSC data.\n- **Caps.** The flow refuses with a localized message that names the\n metric (`content_analyses`) — buy a pack or wait for the reset.\n- **Provider degradation.** Firecrawl or an AI provider outage surfaces\n as `partial` or `refunded` — never as a silent success.\n\n## Connecting the MCP surface\n\nIf you use Claude Code, Claude Desktop, Cursor, or VS Code, you can drive\nContent Intelligence from the RankMeFast MCP endpoint — see\n[RankMeFast MCP](./rankmefast-mcp.en.md).\n\n## Export and deletion\n\nEvery analysis is included in the account export. Deleting your account\npurges analyses inside the grace window documented in\n[Account security](./settings-security.en.md). We never publish anything\nyou have not explicitly accepted."},{"slug":"ai-visibility","locale":"en","title":"AI Visibility","description":"Track whether AI assistants mention your brand. Agency feature.","section":"audits","order":4,"searchText":"# AI Visibility\n\nAI assistants — the chat-style answers people increasingly get instead of a page of blue links — now decide what many buyers see first. AI Visibility tracks whether those answers mention your brand, point people to your site, or send them to a competitor. It is an **Agency** feature.\n\n## Tracked prompts\n\nA prompt is a question a buyer might ask an AI assistant — for example, \"best project management tool for a small team\". You add the prompts that matter to your business. You can track up to **10 prompts per site**. Pick the ones closest to a buying decision; that is where being mentioned is worth the most.\n\n## Running a check\n\nA check runs each tracked prompt and reads the answer the assistant gives back. For every prompt it records:\n\n- whether your brand is **mentioned** by name,\n- whether your site is **linked** as a source,\n- which **competitors** are named,\n- whether the mention is **positive, neutral, or negative**.\n\n\"Mentioned\" simply means your brand name appears in the answer. Being linked as a source is tracked separately and is the stronger signal.\n\n## Share of voice\n\nAcross all your tracked prompts, share of voice is the share of answers that mention you at all. Rising share of voice means you are showing up in more of the answers buyers see; a competitor climbing while you stay flat is an early warning.\n\n## An honest note\n\nThis is **opportunity tracking, not a ranking guarantee**. AI answers are generated fresh each time and can change between runs, between people, and between regions — the same prompt can mention you today and not tomorrow. Treat the trend across many checks as the signal, never a single result. Nothing here lets us, or anyone, promise a place in an AI answer.\n\n## Limits\n\nTracked prompts are capped at 10 per site, and each account has a monthly allowance of checks. When the allowance runs out, checks pause until the next monthly reset — your existing history stays visible. See [Plans, limits & credits](./plans-limits-credits.en.md) for the numbers.\n\n[Back to the docs index](./index.en.md)"},{"slug":"local-seo","locale":"en","title":"Local SEO","description":"Business listing accuracy, reviews, and local map rank. Pro feature.","section":"audits","order":5,"searchText":"# Local SEO\n\nIf customers find you by searching for a place — \"plumber near me\", \"coffee shop downtown\" — most of that traffic comes from the map results and business listings, not the regular blue links. Local SEO tracks the three things that decide whether you show up there. It is available on **Pro** and **Agency**.\n\n## Listing accuracy\n\nSearch engines trust a business whose details match everywhere they appear. RankMeFast checks that your **business name, address, and phone number** are consistent across the listings we can see. Mismatches — an old address, a second phone number, a shortened name — quietly push you down in local results. We flag each inconsistency so you can fix it at the source.\n\n## Reviews and questions\n\nWe track how many **reviews** you have, your **average rating**, and how fresh the reviews are, plus the public **questions and answers** on your listing. Steady, recent, well-answered reviews are one of the strongest local signals; a stall or a run of low ratings shows up here early.\n\n## Local map rank\n\nFor the keywords you already track, we also record where you appear in the **local map results** — the small pack of businesses shown above the regular results. This is separate from your normal search position: a site can rank well in ordinary results and still be missing from the map.\n\n## What one check costs\n\nOne **local check** refreshes your listing details, reviews, and questions together, counted as a single unit against your plan. Local map-rank checks are not separate — they use your normal rank-check budget. When your monthly local checks run out, they pause until the next reset, or you can top up with credits. See [Plans, limits & credits](./plans-limits-credits.en.md) for the exact numbers.\n\n[Back to the docs index](./index.en.md)"},{"slug":"backlinks-competitors","locale":"en","title":"Backlinks & competitors","description":"Domain rating, link rows, competitor overlap.","section":"audits","order":6,"searchText":"# Backlinks & competitors\n\nBacklinks are links from other sites to yours. Google treats them as a vote of confidence — the stronger the linking sites, the stronger your own site looks. Competitor analysis compares your link profile to other sites in your space.\n\n## Domain rating\nDomain rating is a 0–100 score of how strong a site’s **link profile** looks: the count and authority of sites linking to it. **Not** a measure of traffic or content quality. Ballpark:\n- 0–20 — small or new\n- 20–50 — established\n- 50–80 — strong\n- 80–100 — very few sites reach this\n\nTwo different providers score the same site slightly differently. Compare within the same tool over time, not across tools.\n\n## Row limits\nThe backlink list can be enormous — sites with a lot of history can have millions of rows. Each plan caps how many rows we pull per month:\n- Pro — up to a monthly cap on total rows.\n- Agency — unlimited rows.\n- Starter — no backlink data.\n\nSee [Plans, limits & credits](./plans-limits-credits.en.md) for the exact numbers.\n\n## Competitor overlap and the gap view\nAdd competitor domains and RankMeFast shows the **overlap** (keywords/backlinks you both target) and the **gap** (keywords they rank for that you do not). This is the fastest way to find low-hanging content ideas. Available on **Agency**.\n\nNote: numbers reflect what the provider has crawled. Recently launched or heavily paywalled sites can look weaker than they are.\n\n[Back to the docs index](./index.en.md)"},{"slug":"weekly-pulse","locale":"en","title":"Weekly Pulse","description":"One deterministic weekly digest per site, from signals you have already paid for.","section":"audits","order":7,"searchText":"# Weekly Pulse\n\nWeekly Pulse is an opt-in email digest that summarizes what changed for your site over the last ISO week. It is off by default. Each verified team member turns it on for themselves — nobody enables it for you.\n\n## What one pulse costs\n\nEmitting a digest for a site consumes **exactly one `ai_mentions_checks` unit** for that site, no matter how many teammates are subscribed. One collection, many recipients. Reading history, re-opening a past digest, or exporting stored data costs nothing and calls no provider.\n\n## What it contains\n\nThe digest is deterministic — it is rendered from signals we already stored:\n\n- New and lost citations across the AI engines your plan covers.\n- Confirmed rank drops from your tracked keywords.\n- Actions completed or regressed since the prior compatible pulse.\n- Your next three open actions.\n- Google Search Console generative-AI appearance, when Google returned rows.\n\nThe pulse never runs a fresh AI answer, and never issues a new SERP request.\n\n## When it runs\n\nEach site has a fixed weekly slot, deterministically derived from the site ID and spread across Monday–Saturday, 09:00–14:00 UTC. Sunday is reserved. You can see the next slot on the AI Visibility workspace.\n\n## States you may see\n\n- **Completed** — every supported engine returned data and the digest is ready.\n- **Partial** — some engines returned partial data; the digest still ships with what we have.\n- **Unsupported** — none of your plan's engines support this site's market and cohort; no unit is consumed.\n- **Blocked** — your `ai_mentions_checks` allowance is exhausted for the month; no provider call is made.\n- **Failed** — every supported engine failed for this run. The unit is consumed once and never refunded.\n\n## Turning it on or off\n\nOpen the site's AI Visibility workspace, review the spend preview, then toggle **Weekly Pulse**. Disabling suppresses future emails immediately. Re-enabling picks up the next weekly slot.\n\n## Privacy\n\nEmails contain bounded summaries only — host, count, keyword text, rank number, action verb, appearance name. Raw AI answers, source excerpts, competitor prose, prompts, and vendor task IDs never appear in the digest.\n\n[Back to the docs index](./index.en.md)"},{"slug":"confirmed-rank-alerts","locale":"en","title":"Confirmed rank alerts","description":"How a rank change becomes confirmed, and what the observation labels mean.","section":"audits","order":8,"searchText":"# Confirmed rank alerts\n\nA drop only becomes an alert after we observe it twice. That keeps noisy one-off movements out of your inbox.\n\n## How a change becomes confirmed\n\n1. A scheduled rank check finds that one of your top keywords fell by several positions.\n2. We run one fresh second observation for the same keyword in the same market — never a cached result.\n3. The second observation still shows the drop? Then it is a **confirmed rank change**, and only then do we alert you.\n\n## Volatile and unconfirmed changes\n\n- **Volatile** — the second observation shows the keyword recovered. We keep the record as history and send nothing.\n- **Unconfirmed** — we could not complete the second observation (monthly allowance exhausted, or the provider failed). It stays informational and never alerts.\n\nAll three outcomes remain visible next to the keyword's rank history, so nothing is hidden.\n\n## Observation labels\n\nEvery entry states exactly what we observed and when:\n\n- **Market** — the country, language, and device the observation used.\n- **Window** — the time of the first and the second observation.\n- **Freshness of the observation** — how current the data is. A stale observation is labeled as stale, never presented as real-time.\n- **Source** — which provider surface produced the observation.\n\n## Where alerts go\n\nA confirmed rank change appears in [Next Actions](./next-actions.en.md) with a direct link to the keyword's drop record. Dismiss or reopen it there — every decision is kept in the history.\n\n## Reading is free\n\nOpening the alert list, a drop record, or the history reads stored data only. A read never triggers a fresh SERP request and never consumes a unit from your plan.\n\n[Back to the docs index](./index.en.md)"},{"slug":"next-actions","locale":"en","title":"Next Actions","description":"One prioritized list of what to do next, with a full decision history.","section":"audits","order":9,"searchText":"# Next Actions\n\nNext Actions gathers everything worth doing into one prioritized list per site. No hunting through separate reports.\n\n## Where items come from\n\n- Audit findings from your latest finished audit (Fix now and Watch).\n- Confirmed rank changes from your tracked keywords.\n- Search Console and analytics declines.\n- Content recommendations and citation gaps from Content Intelligence.\n- Accepted signals from Audience Research.\n\n## How the list is ordered\n\nThe order is deterministic — the same evidence always produces the same list, and there is no hidden score. We sort by severity, then first-party impact, then confidence (high / medium / low), then how many pages are affected, then effort, then the newest observation; ties break on a stable identifier so the order never jumps.\n\n## Deciding on an item\n\nFour verbs: **Plan**, **Dismiss**, **Complete**, **Reopen**. Every decision is appended to the item's history — nothing is overwritten or deleted.\n\n## Dismissing a finding never edits the audit\n\nDismiss an audit finding as a false positive and it leaves the list. The audit report itself stays untouched — the finding remains in the report exactly as observed.\n\n## Retest\n\nRetest exists for audit findings only. It shows a spend preview first; running it uses one audit from your monthly allowance, the same as starting an audit yourself.\n\n## Reads are free\n\nOpening the list or an item's history reads stored data only. It never calls a provider and never consumes a unit.\n\n[Back to the docs index](./index.en.md)"},{"slug":"ai-visibility-citations","locale":"en","title":"AI citations & source gaps","description":"Sampled prompt checks — which sources AI answers cite, and where you are missing.","section":"audits","order":10,"searchText":"# AI citations & source gaps\n\nAI Visibility checks what AI assistants answer for the prompts you track, and which sources those answers cite. This page explains citations, source gaps, and the states around them.\n\n## A sample, never full coverage\n\nYou track a small set of prompts per site (up to ten). One check asks those prompts across the engines your plan covers, so the result is a sample — sample size (prompts × cohort) — of what people might see. It is never full coverage of everything people ask an AI.\n\n## Supported engines and surfaces\n\nMention checks read the provider's mentions index for Google and ChatGPT surfaces. Answer checks cover ChatGPT, Gemini, and Claude; Perplexity supports live answers only. New surfaces reported by the provider pass through under their own name — we never relabel them.\n\n## Citations\n\nA citation is a URL plus the identity of the source the answer pointed at. We record the exact cited URL whenever the engine reports one, and whether an answer cited you or a tracked competitor.\n\n## Source gaps\n\nA source gap means answers in your topic cite sources — but never yours. Source gaps surface as items in [Next Actions](./next-actions.en.md), each pointing at the sources the engines preferred.\n\n## States\n\n- **Partial results** — some engines returned data, others did not. We show what arrived and name what is missing.\n- **Unsupported for this engine / market** — the engine cannot run this check where you are. It is excluded, never shown as a zero.\n- **Unavailable (not zero)** — no data came back. `unavailable ≠ 0`: it means we do not know, not that you were never cited.\n\n## Not AI-market share\n\nShare of voice compares how often answers name you versus the competitors you track — inside your sampled prompts only. It is not a market-share measurement, and we never present it as one.\n\n## What a check costs\n\nOne check uses one `ai_mentions_checks` unit per tracked prompt. Reading stored results, citations, and source gaps is free.\n\n[Back to the docs index](./index.en.md)"},{"slug":"rank-tracking","locale":"en","title":"Rank tracking","description":"Adding keywords, reading positions, weekly vs daily.","section":"research","order":1,"searchText":"# Rank tracking\n\nRank tracking watches where your pages appear on Google for the keywords that matter to you.\n\n## Adding keywords\nOpen the site view and use **Add keyword**. Type the phrase you want to track. RankMeFast checks it against your keyword cap (see [Plans, limits & credits](./plans-limits-credits.en.md)) and refuses duplicates.\n\n## Reading positions\n- **Position** — the rank of your best-ranking page for that keyword. Position 1 means the top blue link. \"Not in top 100\" means we searched but did not find you in the first hundred results.\n- **Delta** — the change since last check. **-3** means you moved up three positions; **+5** means you slipped five.\n- **Last checked** — how recently the position was measured. Weekly checks land once a week; daily checks land every day.\n\n## Weekly vs daily\nEvery plan gets **weekly** checks bundled. **Daily** checks are a paid add-on — they are more expensive to run because they hit Google’s SERP much more often.\n\n## Trend chart\nSelecting a keyword shows a trend of your position over time. A rising line is bad (higher number = worse rank); a falling line is good. Under the chart is an accessible data table with the same numbers.\n\n## \"Unavailable\" checks\nSometimes a check comes back as unavailable. The most common reasons:\n- Our SERP provider’s quota is exhausted for the day.\n- The keyword returns a SERP feature we do not measure (news carousel only, for instance).\n- Google served a captcha and we could not verify the result.\nUnavailable checks do not consume credit and are retried on the next scheduled run.\n\n## Cap reached\nEvery plan has a keyword cap. When you hit it, the Add form is replaced with a notice pointing you to upgrade. Existing keywords keep running.\n\n[Back to the docs index](./index.en.md)"},{"slug":"keyword-research","locale":"en","title":"Keyword research","description":"Volume and difficulty in plain words. Pro feature.","section":"research","order":2,"searchText":"# Keyword research\n\nKeyword research helps you decide which phrases to target. It is a **Pro** and **Agency** feature — Starter accounts can upgrade to unlock it (see [Plans, limits & credits](./plans-limits-credits.en.md)).\n\n## Volume\nSearch volume is an estimate of how often a phrase is searched for on Google, per month, in the country you pick. It is an average smoothed over recent months — not a live counter. Values under 20 are noisy; treat them as directional.\n\n## Difficulty\nDifficulty is a 0–100 score of how hard it looks to rank on page 1 for the phrase. Higher = harder. The score blends how strong the current top-10 sites look and how many quality backlinks they have. A difficulty of 25 on a well-established site is easy; the same 25 on a brand-new domain is not.\n\n## Caching\nVolume and difficulty do not change hour-to-hour — so RankMeFast caches the answer for a while. If two people on your account look up the same phrase in the same country in the same week, the second request comes back instantly and does not count against your metered usage. If you re-run after the cache expires, we go back to the provider.\n\n## Suggestions and questions\nWhen available, we also show related phrases and question variants. Both use the same volume/difficulty numbers.\n\nNote: numbers come from a third-party provider and are rounded. Two different tools rarely agree exactly. Trust the ranking of one number against another more than the absolute value.\n\n[Back to the docs index](./index.en.md)"},{"slug":"google-search-console","locale":"en","title":"Connecting Google Search Console","description":"Why connect, what we access, how to disconnect.","section":"research","order":3,"searchText":"# Connecting Google Search Console\n\nSearch Console is Google’s free tool that shows how Google sees your site. Connecting it to RankMeFast lets three audit rules use the real answer instead of guessing:\n- **Not indexed by Google** — is a specific page actually in search?\n- **Rich results issues** — are your structured-data snippets valid?\n- **Indexed with warnings** — has Google flagged a canonicalization or duplicate signal?\n\n## Why connect\nWithout a connection, these rules stay in **Watch** with \"insufficient data\". With a connection, they land in the right bucket with actionable pages.\n\n## What we access (and what we do not)\n- We ask Google for the **read-only** Search Console scope: `webmasters.readonly`. That is the smallest scope that lets us call the URL Inspection API.\n- We **never** get access to Gmail, Drive, or your Google account beyond Search Console.\n- We store an encrypted refresh token so we can renew the short-lived access token. Nothing else.\n\n## Connecting\n1. Open Settings → Google Search Console.\n2. Click **Connect Google Search Console**.\n3. Sign in to Google and grant the read-only scope.\n4. RankMeFast matches the connected property to your site.\n\n## Disconnecting or revoking\nClick **Disconnect** in the same panel and confirm. That deletes our stored token immediately. You can also revoke us at any time from https://myaccount.google.com/permissions.\n\n## \"Reconnect needed\"\nIf Google invalidates our token — for instance, when you change your account password — the card shows a red **Reconnect needed** banner. The audit rules fall back to \"insufficient data\" until you reconnect.\n\n[Back to the docs index](./index.en.md)"},{"slug":"gsc-generative-appearance","locale":"en","title":"Generative AI appearance in Search Console","description":"A first-party GSC signal — how often your site was shown inside Google generative AI features.","section":"research","order":4,"searchText":"# Generative AI appearance in Search Console\n\nGoogle Search Console reports how often your pages appeared inside Google's generative-AI surfaces (for example, AI Overviews). This card shows those numbers verbatim — we do not add, invent, or estimate anything Google did not send.\n\n## What we read\n\nWe call the Search Console API with `dimensions=['searchAppearance']` over the last 28 days ending three days ago (Google's own reporting lag). Every row Google returns is classified from primary Google documentation:\n\n- Known generative-AI values map to a stable label.\n- Unknown values pass through as `Other` with the raw name preserved for support review — never silently reclassified as generative.\n\n## States\n\n- **Available** — Google returned at least one recognized generative-AI row. The card shows the exact clicks, impressions, CTR, and average position Google sent, per appearance row.\n- **Unavailable** — Google returned no recognized generative-AI row for this property and window. This is different from a zero — it means Google did not report the surface at all. We do not zero-fill.\n- **Partial** — Google returned rows but flagged the response as partial (rate limit, truncation). The rows we did receive are still displayed.\n- **Reconnect required** — the Google connection needs to be reconnected. Open the site's Google workspace and follow the reconnect prompt.\n\n## Separate from provider metrics\n\nGenerative-AI appearance is a **first-party** Google signal. It is displayed on the Google workspace next to your other Search Console cards. It is never merged with the mention and share-of-voice charts on the AI Visibility workspace — those come from a different provider and measure a different thing.\n\n## What this does not tell you\n\nSearch Console reports impressions and clicks — how often Google chose to include your page inside a generative surface. It does not tell you whether the AI answer named you, linked you, quoted you, or preferred a competitor. For that, use [AI Visibility](./ai-visibility.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"audience-research","locale":"en","title":"Audience Research","description":"Turn topics, competitors, and a market into audience segments from public evidence.","section":"research","order":5,"searchText":"# Audience Research\n\nAudience Research turns your market, a few seed topics, and optional competitor domains into audience segments — built only from public evidence.\n\n## What you provide\n\n- Your site's market (country and language).\n- Up to ten seed topics you want to explore.\n- Up to five competitor domains (optional).\n\n## Preview before you spend\n\nThe preview shows what a run will use and how many runs you have left this month. It reserves nothing — previewing is always free.\n\n## What a run costs\n\nOne run consumes one `audience_research_runs` unit, no matter how many pages it reads inside. Monthly runs per plan: Free 0, Starter 2, Pro 10, Agency 30. There is no credit pack for this metric — at the cap, wait for the monthly reset or upgrade.\n\nStarting the identical run twice does not charge twice — the second start returns the run already in flight. A later fresh run is a new run and uses a new unit.\n\n## A fixed per-run cost ceiling\n\nEvery run has a fixed per-run cost ceiling, with a stricter inner ceiling for the AI clustering step. A run that reaches the ceiling stops early and keeps the evidence it already collected — the result is labeled **partial results**, never silently truncated.\n\n## How pages are selected\n\nSelection is deterministic — the same inputs pick the same public pages. We rank candidates by topic relevance, then search position, then recency, and spread picks across source types and domains so no single site can dominate the evidence.\n\n## Run states\n\nA run moves through **queued → discovering → selecting → collecting → clustering** and ends as **completed**, **partial results**, or **failed**.\n\n## Refunds\n\nA unit is returned exactly once, and only when a run fails with no usable public evidence at all. A completed or partial run keeps its unit — partial evidence is still evidence.\n\n## Nothing is published\n\nResults are for your account only. A run never publishes, posts, or shares anything anywhere.\n\n## Where results go\n\nSignals you accept appear in [Next Actions](./next-actions.en.md) and can create Content Intelligence briefs. Nothing moves without your decision.\n\n[Back to the docs index](./index.en.md)"},{"slug":"keyword-intelligence","locale":"en","title":"Keyword Intelligence","description":"Gap, overview, trends, and clustering — what each check costs and what the labels mean.","section":"research","order":6,"searchText":"# Keyword Intelligence\n\nKeyword Intelligence is the workspace for keyword gaps, overviews, trends, and AI clustering. It is available on Pro and Agency plans.\n\n## What each check costs\n\n- **Gap** — compares your domain against competitors. One `keyword_lookups` unit per competitor in the comparison (duplicates are removed first).\n- **Overview** — one `keyword_lookups` unit per keyword you look up.\n- **Trends** — one `keyword_lookups` unit per keyword.\n\nA result served from cache still counts as one unit. Caching keeps vendor costs (and your price) down; a check is still a check.\n\n## Reading the labels\n\nEvery number carries a label saying where it came from:\n\n- **Estimate** — derived by simple arithmetic from provider data, such as a trend line computed from monthly volumes.\n- **Provider observation** — a value the data provider reported; we display it as received.\n- **AI interpretation** — produced by an AI step, such as cluster names. Treat it as a suggestion, not a measurement.\n\n## Clustering\n\nClustering groups a keyword list into topics with one AI step. One clustering run uses one `ai_summaries` unit; rerunning the identical list returns the stored result for free.\n\n## When keyword lookups run out\n\nThe `keywords-1k` credit pack adds 80 `keyword_lookups` for $19, one-time. It tops up gap, overview, and trends reads after your monthly allowance is spent — see [Plans, limits & credits](./plans-limits-credits.en.md).\n\n## From clusters to work\n\nAccept a cluster and it routes onward — into a Content Intelligence brief or into your SEO work, following the route suggested on the cluster. Dismissing records the decision and changes nothing else. Every decision is kept.\n\n[Back to the docs index](./index.en.md)"},{"slug":"link-intelligence","locale":"en","title":"Link Intelligence","description":"Run deeper backlink checks and compare link gaps without guessing what a unit covers.","section":"research","order":7,"searchText":"# Link Intelligence\n\nUse Link Intelligence when the standard backlink overview is not detailed enough. It can pull referring domains, anchors, history, or bulk domain ranks, and it can compare your site with up to three competitors.\n\n\n\n## Know what one unit covers\n\n- One submitted deep pull uses one `link_intel_checks` unit.\n- A link-gap run uses one unit for each competitor. Comparing three competitors uses three units.\n- A preview shows the exact unit breakdown but spends nothing. Confirming the run checks your allowance again.\n\nSee [Pricing](./pricing.en.md) for current plan allowances and pack prices. The deep tools require an eligible plan; opening a result you already created is a stored-data read.\n\n## Understand cache and refunds\n\nA result served from cache still counts. A successful check with no rows also counts: it confirms that the source returned no matching links at that time.\n\nIf the source fails before RankMeFast retains any row, that failed unit is returned once. In a multi-competitor gap run, each competitor is settled separately, so useful legs remain charged while a zero-evidence failed leg is refunded. A preview, cancelled preview, or stored-result read never uses a unit.\n\n## Read the result honestly\n\nDeep results carry the time they were observed. A missing first-seen date stays unavailable rather than being invented. Gap percentages are computed from the stored rows, and competitor input order is preserved.\n\nThe checks are **on demand**, not continuous monitoring. Run a new check when you need a fresher observation. Turning the feature off stops new previews and runs, but stored results remain readable.\n\n## Start a check\n\nOpen a site's Backlinks workspace. Choose **Domains**, **Anchors**, **History**, or **Gap**, review the spend preview, then confirm. Use the stored-runs list to reopen a prior result without spending again.\n\nFor monthly caps, packs, and the common refund rule, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"traffic-insights","locale":"en","title":"Traffic Insights","description":"Compare modeled competitor traffic while keeping estimates, units, and refunds clear.","section":"research","order":8,"searchText":"# Traffic Insights\n\nTraffic Insights creates a stored snapshot of a competitor domain: modeled monthly organic visits, domain rank, keyword count, leading countries, and history. Use it for directional comparison, not as a replacement for the competitor's analytics.\n\n\n\n## Know what one unit covers\n\nOne submitted target-domain snapshot uses one `traffic_snapshots` unit. A multi-domain preview shows one unit for each domain, but each confirmed domain becomes its own snapshot. Previewing or cancelling the preview spends nothing.\n\nCurrent monthly allowances and pack prices are on [Pricing](./pricing.en.md). Opening, filtering, or comparing snapshots you already created is a stored-data read and uses no new unit.\n\n## Treat every number as an estimate\n\nEvery numeric Traffic Insights field is labeled **Estimate**. The values are modeled from a search-data index; they are not first-party visits, conversions, or analytics. Missing values remain unavailable. A partial snapshot can still be useful when one part succeeds, and the interface shows that it is partial.\n\nCompare views align up to five stored snapshots. They do not fill gaps or infer missing country or history values.\n\n## Understand cache and refunds\n\nA cached snapshot still counts as one unit. A successful empty result also counts.\n\nThe snapshot collects several related observations. If all of them fail and RankMeFast retains no result, the unit is returned exactly once. If any usable observation is retained, the snapshot is stored as partial or complete and the unit remains consumed. A transport retry cannot return the same unit twice.\n\n## Choose when to refresh\n\nTraffic Insights runs **on demand**. It does not monitor competitors continuously. Start another snapshot when you need a newer estimate. If an operator pauses new snapshots, your stored list, detail, and comparison views remain readable.\n\nOpen a site's **Traffic** tab, enter a public domain, review the spend preview, and confirm. For caps, packs, and refund rules, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"keyword-trends","locale":"en","title":"Keyword Trends","description":"Explore search-interest direction, momentum, and seasonality with honest estimate labels.","section":"research","order":9,"searchText":"# Keyword Trends\n\nKeyword Trends shows how search interest changes over time. It reports a relative index, year-over-year direction, recent momentum, and seasonal months for up to five keywords in one exploration.\n\n\n\n## Know what one unit covers\n\nOne confirmed exploration uses one `trend_explorations` unit, whether you enter one keyword or five. Previewing the request, cancelling the preview, reopening a stored run, or selecting a related-query follow-up before confirming does not spend a unit. A follow-up you confirm is a new exploration.\n\nSee [Pricing](./pricing.en.md) for current allowances and pack prices.\n\n## Read the index as an estimate\n\nEvery series and readout is labeled **Estimate**. The 0–100 search-interest index is relative to the selected market and time window. It is not absolute monthly search volume, traffic, or a forecast.\n\n- Year-over-year needs at least 56 valid observations.\n- Momentum needs 12 observations and can be up, down, or flat.\n- Seasonality needs 24 distinct months. Without enough history, RankMeFast says so instead of inventing a pattern.\n\nA flat, sparse, or empty successful response is a valid observation, not a failure.\n\n## Understand cache and refunds\n\nA result served from cache still uses one unit. A successful exploration with no series also uses one unit.\n\nIf the source fails before any series is retained, the reserved unit is returned exactly once and the run is marked refunded. Once a usable series is retained, the unit remains consumed even if another readout is unavailable.\n\n## Choose when to explore\n\nKeyword Trends is **on demand**, not continuous monitoring. Run it again when you need a newer view. If new explorations are paused, stored runs stay readable.\n\nOpen **Live trends** from Keyword Research, enter up to five phrases, choose the market and language, review the spend preview, and confirm. For monthly caps and the common refund rule, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"review-intelligence","locale":"en","title":"Review Intelligence","description":"Sync public reviews on demand, read trends and cited themes, and understand exactly when a unit is returned.","section":"research","order":10,"searchText":"# Review Intelligence\n\nReview Intelligence brings public reviews from configured Google, Trustpilot, and Tripadvisor listings into one stored inventory. It calculates rating statistics and can group repeated praise and complaints into themes supported by stored review citations.\n\n\n\n## Know what one unit covers\n\nOne confirmed sync uses one `review_syncs` unit for the whole job. You may select one, two, or three configured sources and choose a depth up to 100 reviews per source; the unit cost does not multiply by source count. The theme pass is included.\n\nPreviewing or cancelling spends nothing. See [Pricing](./pricing.en.md) for current monthly allowances and pack prices.\n\n## Understand cache, deduplication, and refunds\n\nPublic review material served from cache still counts. RankMeFast deduplicates by source and source review ID, so a successful sync that finds no new rows also counts: the source was checked and existing reviews were confirmed.\n\nThe unit is returned exactly once only when **every selected source** ends in a source failure and no new review row is retained. A partial sync, an empty or fully deduplicated success, or a theme-generation failure still consumes the unit because review work succeeded or evidence remains. Retrying cannot return the same unit twice.\n\n## Read themes and statistics\n\nRatings, counts, source splits, and monthly trends come from stored review rows. Generated themes are optional. A theme is shown only when it cites stored reviews; unreliable theme text is dropped. If theme generation fails, your reviews and deterministic statistics remain available.\n\nSearch, filter, open a stored run, and export matching rows as CSV without using another sync unit. Review text is displayed as text, not executable markup.\n\n## Choose when to sync\n\nReview Intelligence is **on demand**. It does not monitor listings continuously, send review alerts, reply to reviews, or draft responses. Start a new sync when you want a fresher inventory. Pausing new syncs does not hide stored sources, runs, reviews, statistics, themes, or exports.\n\nOpen a site's **Reviews** tab, configure at least one source, review the spend preview, and confirm. For caps and refunds, see [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"brand-radar","locale":"en","title":"Brand Radar","description":"Scan one brand query on demand, inspect retained mentions, and understand the add-on and refund boundary.","section":"research","order":11,"searchText":"# Brand Radar\n\nBrand Radar searches for public mentions of one brand query, stores the retained rows, and builds counts, sentiment distribution, top domains, and a comparison with the previous settled scan for the same query. It can also create a short cited digest.\n\n\n\n## Know what one unit covers\n\nOne confirmed scan of one brand query uses one `brand_mention_scans` unit. The optional language/location filters and the cited digest are included. A scan cannot bundle competitor queries into the same unit.\n\nPreviewing or cancelling spends nothing. The preview always says a fresh fetch is required because mention rows are account-scoped and are not shared through a cross-account cache. Current base allowances, the optional Brand Radar add-on, and pack prices are on [Pricing](./pricing.en.md).\n\n## Understand refunds and partial results\n\nThe unit is returned exactly once only if the initial mention search ends in a source failure and RankMeFast retains no mention row.\n\nThese outcomes still consume the unit:\n\n- the search succeeds but finds no mentions;\n- some mentions are retained before a later source or budget stop;\n- the summary or generated digest fails;\n- every generated digest sentence is dropped because its citation cannot be verified.\n\nRetained evidence is never discarded just to create a refund.\n\n## Separate measured facts from generated prose\n\nMention count, sentiment split, top domains, and change from the previous same-query scan are computed from stored rows. No previous scan means “no comparison yet,” never a made-up zero.\n\nDigest sentences are generated prose and are shown only with citations to retained mention rows. If no reliable sentence survives, the interface says so. You can inspect mention rows and export them as CSV.\n\n## Choose when to scan\n\nBrand Radar is **on demand**, not continuous monitoring. Start a new scan when you need a new observation. Weekly Pulse may summarize deltas already stored by Brand Radar, but it does not start a scan or spend another scan unit. If new scans are paused, stored lists, details, citations, trends, and exports remain readable.\n\nFor eligibility, allowances, and refund rules, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"serp-sensor","locale":"en","title":"SERP Sensor methodology","description":"How the public daily volatility index is collected, calculated, budgeted, and limited.","section":"research","order":12,"searchText":"# SERP Sensor methodology\n\nThe [SERP Sensor](/serp-sensor) is a shared public index of observed search-result movement. It is not a personal rank tracker, a forecast, or a monitor for your own keyword set.\n\n\n\n## Fixed basket and cadence\n\nRankMeFast checks the same 200 fixed US-English keywords once per UTC day: 20 keywords in each of 10 categories. For every keyword it keeps the first 20 unique public HTTP or HTTPS result URLs. Users cannot change the keywords, market, language, depth, or collection time.\n\nThe categories cover finance, health, travel, technology, retail, entertainment, local services, education, real estate, and automotive. The fixed basket makes one day comparable with the next; it does not represent every search or location.\n\n## How the index is calculated\n\nFor one keyword, each result's current rank is compared with its previous rank. A result missing on one day is treated as rank 21. The absolute rank movements across the union of both result sets are added and divided by 420, the maximum displacement for a top-20 set. This produces a value from 0 to 1.\n\nRankMeFast averages the 20 keyword values in a category and scales the result to 0–10, rounded to one decimal place. The composite is the average of the category values that were published that day. The labels describe observed displacement:\n\n| Index | Label |\n|---|---|\n| 0.0–2.0 | Low |\n| above 2.0–4.5 | Routine |\n| above 4.5–7.0 | Elevated |\n| above 7.0–10.0 | High |\n\nThe first complete observation is a **baseline** and has no numeric change. If only some categories complete, the composite is marked partial. Missing data is not estimated or filled.\n\n## Budget and availability\n\nThe sensor uses a global operator budget, not a customer allowance. It consumes no account unit, has no customer cache rule, and has no refund because customers are never charged. The collector checks its budget before and during the daily run. A category skipped for budget or collection failure is labeled accordingly.\n\nPublic history contains at most the latest 30 stored dates per category. Turning the sensor off returns an unavailable state without exposing the fixed keywords or result URLs.\n\nFor customer-run research features and their units, see [Pricing](./pricing.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"pricing","locale":"en","title":"Pricing","description":"How RankMeFast plans and monthly allowances work.","section":"account","order":1,"searchText":"# Pricing\n\nRankMeFast sells you a monthly allowance of SEO checks. Free forever tier for kicking the tyres; Starter and Pro for founders and teams; Agency for client work.\n\n## What we count\n\nEvery plan meters the same units:\n\n- **Audit** — one full audit run over a site.\n- **Audit page** — one page inside that run (crawl budget).\n- **Tracked keyword** — one keyword you keep watching.\n- **Rank check** — one SERP fetch for a tracked keyword; the weekly sweep and every manual re-check count.\n- **Keyword lookup** — one site-keyword discovery request or one read from the full keyword-research panel.\n- **Backlink row** — one row on the backlink list.\n- **AI summary** — one Claude-written plain-language recap of the fix-now list.\n- **Competitor lookup** — one competitor list or intersection call.\n\nCached results still count as one check. Caching is how we keep vendor costs (and your price) down; a check is still a check.\n\nMonthly keyword lookups are **5 Free / 20 Starter / 50 Pro / 100 Agency**. Site discovery is included on every plan; full keyword research remains Pro and Agency.\n\n## New intelligence units\n\n\n- **Keyword Trends exploration** — one confirmed exploration of up to five phrases.\n- **Traffic Insights snapshot** — one confirmed target domain.\n- **Link Intelligence check** — one deep pull, or one unit for each competitor in a link-gap run.\n- **Review Intelligence sync** — one whole sync of one to three configured sources.\n- **Brand Radar scan** — one brand query; its cited digest is included.\n\nThese five features run on demand, not as continuous monitoring. Cache hits count, and successful empty results count. A reserved unit is returned exactly once only when the feature's source fails before it retains any usable evidence; each guide explains the narrower multi-part rule. Brand Radar always makes a fresh account-scoped fetch rather than sharing mention rows between accounts.\n\nRead the guides for [Link Intelligence](./link-intelligence.en.md), [Traffic Insights](./traffic-insights.en.md), [Keyword Trends](./keyword-trends.en.md), [Review Intelligence](./review-intelligence.en.md), and [Brand Radar](./brand-radar.en.md).\n\n## Brand Radar add-on and intelligence packs\n\n\n\nThe optional Brand Radar add-on is available on Pro and Agency. It adds 60 scans each month for $19/month. Agency already includes 20 monthly scans; Pro needs the add-on to start scanning.\n\nOne-time intelligence packs stay on your balance until used:\n\n- 40 Brand Radar scans for $29.\n- 50 Link Intelligence checks for $19.\n- 50 Review Intelligence syncs for $19.\n- 100 Traffic Insights snapshots for $19.\n- 200 Keyword Trends explorations for $19.\n\n## What happens at a cap\n\n1. The app warns you before you hit a cap.\n2. New requests hard-stop at the cap. Nothing is billed beyond the ceiling.\n3. Three ways out: wait for the monthly reset, buy a credit pack (+80 keyword lookups for $19, or +5 audits for $29 — one-time, never expire), or upgrade your plan.\n\nThe keyword pack is the `keywords-1k` pack: +80 `keyword_lookups` for $19, one-time. It tops up keyword gap, overview, and trends reads when your monthly allowance runs out. Audience Research runs have no credit pack — upgrade or wait for the reset.\n\n## Yearly billing\n\nYearly gives you two months free — 20% off the monthly price. Switch between monthly and yearly anytime from the billing page.\n\n## Cancelling\n\nCancel from the billing page. No calls, no retention gauntlet. Your audits and rank history remain read-only for 30 days."},{"slug":"plans-limits-credits","locale":"en","title":"Plans, limits & credits","description":"Starter, Pro, Agency — caps, trial, overage credits.","section":"account","order":2,"searchText":"# Plans, limits & credits\n\nRankMeFast has three subscription tiers and a free one-off audit.\n\n## Free\n- One full audit at sign-up. No credit card. No ongoing checks.\n\n## The three tiers\n\n\n| Plan | Monthly (USD) | Yearly (USD) | Sites | Keywords | Audits/month | Audit pages | Backlink rows | AI summaries | Seats | Audience Research runs |\n|----------|---------------|--------------|-------|----------|--------------|-------------|---------------|--------------|-------|------------------------|\n| Starter | $49 | $470 | 2 | 250 | 10 | 1,000 | 0 | 20 | 1 | 2 |\n| Pro | $149 | $1,430 | 5 | 1,000 | 30 | 3,000 | 10,000 | 100 | 3 | 10 |\n| Agency | $479 | $4,598 | Unlimited | 2,000 | 45 | 5,000 | 100,000 | 500 | Unlimited | 30 |\n\nPrices above are in USD. Annual is 20 % off the monthly total. Daily rank tracking is a paid add-on on every plan — never bundled.\n\n## Audience Research runs\n\n\nOne run is one whole research job, no matter how many pages it reads. Free accounts get 0 runs; Starter 2, Pro 10, Agency 30 per month. There is no credit pack for this metric — at the cap, wait for the monthly reset or upgrade. See [Audience Research](./audience-research.en.md).\n\n## Keyword Intelligence and Weekly Pulse units\n\n\n- A keyword gap check uses one `keyword_lookups` unit per competitor you compare.\n- A keyword overview or trends read uses one `keyword_lookups` unit per keyword. Results served from cache still count.\n- Clustering a keyword list uses one `ai_summaries` unit; rerunning the identical list is free.\n- One Weekly Pulse digest uses one `ai_mentions_checks` unit per site per week, no matter how many teammates receive it.\n\nDetails live in [Keyword Intelligence](./keyword-intelligence.en.md) and [Weekly Pulse](./weekly-pulse.en.md).\n\n## Intelligence allowances and units\n\n\n| Plan | Keyword Trends explorations | Traffic snapshots | Link Intelligence checks | Review syncs | Brand Radar scans |\n|---|---:|---:|---:|---:|---:|\n| Free | 0 | 0 | 0 | 0 | 0 |\n| Starter | 10 | 5 | 0 | 0 | 0 |\n| Pro | 40 | 25 | 25 | 10 | 0 |\n| Agency | 80 | 80 | 80 | 50 | 20 |\n\nOne Trends unit covers up to five phrases. One Traffic unit covers one target domain. One Link unit covers one deep pull or one competitor leg in a gap run. One Review unit covers a whole one-to-three-source sync. One Brand Radar unit covers one brand query and its cited digest.\n\nCache hits and successful empty results count. A unit is returned exactly once only when the source fails before the feature retains usable evidence. Multi-part Link, Traffic, and Review jobs apply that rule to their documented legs; see [Link Intelligence](./link-intelligence.en.md), [Traffic Insights](./traffic-insights.en.md), [Keyword Trends](./keyword-trends.en.md), [Review Intelligence](./review-intelligence.en.md), and [Brand Radar](./brand-radar.en.md).\n\n## Brand Radar add-on and intelligence packs\n\n\n\nPro and Agency can add 60 Brand Radar scans per month for $19/month. Agency's 20 base scans remain available without the add-on; Pro's base is zero.\n\nThe one-time packs are 40 Brand Radar scans for $29, 50 Link checks for $19, 50 Review syncs for $19, 100 Traffic snapshots for $19, and 200 Trends explorations for $19. Purchased pack balances remain until used. Current purchase options are always shown on [Pricing](./pricing.en.md) and under **Billing → Credits**.\n\n## 14-day free trial\nEvery paid tier includes a 14-day trial. You can cancel any time during the trial without a charge. Card details are collected at sign-up; the first payment lands on day 15.\n\n## What happens at a cap\nWhen you hit a monthly cap (audits, keywords tracked, backlink rows, AI summaries) the specific action is blocked with a message that points at:\n1. Upgrading to a higher tier.\n2. Buying an **overage credit pack** for the matching metric.\n\nNothing else on the account changes — your existing sites, keywords, and reports stay intact.\n\n## Overage credit packs\nSmall one-time packs of extra usage priced per metric. They top up the matching cap and remain on the account until used. Existing subscribers see them under Billing → Credits.\n\n## Cancelling and downgrading\nCancelling keeps your access until the end of the billing period. Downgrading takes effect at the next renewal — your extra sites and keywords are not deleted, they just become read-only until you delete some or upgrade again.\n\n[Back to the docs index](./index.en.md)"},{"slug":"settings-security","locale":"en","title":"Account security","description":"Change your password and manage account credentials.","section":"account","order":3,"searchText":"# Account security\n\nManage the credentials that protect your RankMeFast account from **Settings → Security** (URL: `/settings/security`).\n\n## Change your password\n\n1. Sign in and open **Settings → Security**.\n2. Fill in your **current password**, then choose a **new password** and confirm it. The new password must be at least eight characters and must differ from the current one.\n3. Click **Update password**.\n\nWe use scrypt to store password hashes — we never see your password in plain text. When you change your password successfully:\n\n- Every **other** active session for your account is signed out. The device you used to change the password stays signed in.\n- We send a confirmation email to the address on your account. If you did not change the password, that email is your signal to reset your password immediately via [the forgot-password flow](./getting-started.en.md).\n\nIf the current password you entered is wrong, we say so inline — we never reveal whether an account exists for a given email.\n\n## Change your email\n\n1. Sign in and open **Settings → Security**.\n2. Enter the **new email** address in the *Change email* card. It must be different from your current one.\n3. Click **Send verification link**.\n\nWe send a confirmation link to the **new** address. Your account email only changes after you follow that link:\n\n- Your **current email keeps working** for sign-in until you confirm.\n- If the new address is already registered to another account, we reject the change with a generic *email unavailable* message — we never reveal who owns that address.\n- The link is single-use and expires after a short window; request another change from the same page if it lapses.\n\n## Forgot your current password?\n\nIf you cannot remember your current password, sign out and use **Forgot password?** on the login screen. The emailed link takes you to a page where you can pick a new password without confirming the old one.\n\n\n## Two-factor authentication\n\nAdd a second step at sign-in so a stolen password alone cannot get into your account.\n\n### Enable two-factor\n\n1. Sign in and open **Settings → Security**.\n2. In the **Two-factor authentication** card, enter your current password and click **Enable two-factor**.\n3. Scan the QR code with an authenticator app like 1Password, Authy, or Google Authenticator — or paste the manual-entry code into the app.\n4. Copy or download the **backup codes**. Each code works exactly once. They are the only way back into your account if you lose your device.\n5. Enter the six-digit code your authenticator app is showing right now and click **Verify and turn on**.\n\nOnce enabled, every sign-in asks for a fresh code from your authenticator (or a one-time backup code).\n\n### Disable two-factor\n\nOpen the same card, click **Disable two-factor**, and enter your current password. We turn the feature off immediately.\n\n### Admin accounts\n\nIf your account has the **Admin** role, two-factor authentication is required. You will not be able to open the admin panel until you enable it from **Settings → Security**.\n\n### Lost your device\n\nUse one of your backup codes on the sign-in challenge screen — the same screen has a **Use a backup code** link. Each code works once; if you run out, an admin on your team can reset the second factor from the admin panel.\n\n## Your data rights\n\nYou can request an export or a deletion of your account data. Both cover everything the research surfaces store for your account: rank observations, AI Visibility citations and source gaps, Audience Research runs and their evidence, and cached keyword results. Deletion respects a grace period before the purge — you can cancel the request during that window."},{"slug":"notification-preferences","locale":"en","title":"Notification preferences","description":"Choose which non-security emails you want to receive.","section":"account","order":4,"searchText":"# Notification preferences\n\nYou control which non-security emails RankMeFast sends you. Open **Settings → Notifications** (`/settings/notifications`) and toggle any of the four switches.\n\nEvery switch defaults to **on**. Turning a switch off takes effect immediately — the next matching email is skipped.\n\n## What each toggle controls\n\n- **Audit complete** — Sent when an audit run finishes for one of your sites.\n- **Rank drops** — Sent when a tracked keyword loses positions in your latest rank check.\n- **Billing** — Payment receipts, failed-payment reminders, and dunning notices. Turning this off silences ALL billing email; you can still see receipts and past invoices in **Billing → Invoices**.\n- **Product updates** — Occasional welcome messages, product news, and tips.\n\n## Weekly Pulse is a separate opt-in\n\nThe [Weekly Pulse](./weekly-pulse.en.md) digest is not controlled by the switches above. Each teammate turns it on per site from the site's AI Visibility workspace, and it stays off until they do.\n\nDelivery retries are safe: a retry can never send the same digest twice. If your address cannot receive the email, other subscribed teammates still get theirs.\n\n## Security email cannot be disabled\n\nThe following emails are always sent, regardless of the switches above:\n\n- Password reset link.\n- Email address verification.\n- Password-changed confirmation.\n\nThose emails exist to protect your account. If you receive one you did not expect, treat it as your signal to review your active sessions and reset your password from **Settings → Security**."},{"slug":"team","locale":"en","title":"Team & invites","description":"Invite teammates, accept invites, and manage seats on your plan.","section":"account","order":5,"searchText":"# Team & invites\n\nEvery RankMeFast account can share access with additional teammates. You manage members and pending invites from **Settings → Team**.\n\n## Seats per plan\n\n| Plan | Included seats |\n|------|----------------|\n| Free | 1 |\n| Starter | 1 |\n| Pro | 3 |\n| Agency | Unlimited |\n\nAn account's owner always counts as one seat. The current usage is shown at the top of the page (for example, \"2 of 3 seats used\"). When every seat is taken, the invite form is disabled and an **Upgrade plan** link appears.\n\n## Invite a teammate\n\n1. Open **Settings → Team**.\n2. Under **Invite a teammate**, enter your colleague's email address.\n3. Click **Send invite**.\n\nWe email a single-use link that is valid for seven days. The pending invite immediately reserves a seat.\n\n## Accept an invite\n\nClick the link in the email — or open `/team/accept/<token>` in your browser while signed in with the invited email. Once accepted, the seat converts from **Pending** to **Active** and the teammate gains access to the workspace.\n\n- If the link is opened by a different account than the one it was addressed to, RankMeFast responds with a \"not found\" error — invite tokens are single-use and can only be redeemed by the invited email.\n- If the link has expired, ask the workspace owner to send a fresh invite.\n- If the invite has already been accepted, opening the link again returns \"already accepted\".\n\n## Remove a teammate or revoke a pending invite\n\n1. Open **Settings → Team**.\n2. Open the ⋯ menu on the member's row.\n3. Choose **Remove** for an accepted member, or **Revoke invite** for a pending one.\n\nRemoving a member (or revoking a pending invite) frees the seat immediately — you can invite someone else without waiting.\n\nThe workspace owner cannot be removed from the team. Team members can remove themselves (a \"leave\" action)."},{"slug":"public-api","locale":"en","title":"Public API","description":"Read your sites, reports, rankings, and keywords over a simple key-authenticated API. Agency feature.","section":"developers","order":1,"searchText":"# Public API\n\nThe public API gives you read-only access to the data RankMeFast already holds for your account — sites, the latest audit report, rank history, and tracked keywords. It is an **Agency** feature (see [Plans, limits & credits](./plans-limits-credits.en.md)). The API never triggers new vendor work — it only reads what your audits and rank checks have already produced.\n\n## Authentication\n\nCreate a key under **Account → API keys** (`/profile?tab=api-keys`). The full key is shown **exactly once** — copy it right away; afterwards only its prefix is visible. You can hold up to ten active keys and revoke any of them at any time. A revoked key stops working immediately.\n\nSend the key as a bearer token on every request:\n\n```\nAuthorization: Bearer rmf_XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX\n```\n\nReplace `https://your-rankme-host` in the examples below with your api origin (the `SERVER_URL` of your install).\n\n## Endpoints\n\n### List your sites\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n https://your-rankme-host/api/v1/sites\n```\n\nReturns `{ \"sites\": [{ \"id\", \"domain\", \"url\", \"createdAt\" }] }`.\n\n### Latest audit report for a site\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n https://your-rankme-host/api/v1/sites/<siteId>/report/latest\n```\n\nReturns the most recent **succeeded** audit as `{ \"runId\", \"report\" }` — the same findings, buckets, and localized copy the dashboard shows. Responds `404` when the site has no finished audit yet.\n\n### Rank history for a site\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n \"https://your-rankme-host/api/v1/sites/<siteId>/rank-history?from=2026-06-01&to=2026-07-01\"\n```\n\nReturns `{ \"keywords\": [{ \"id\", \"phrase\", \"series\": [...] }] }`. Each series point carries the position, the URL that ranked, and the Google AI Overview signals (`aiOverviewPresent`, `aiCited`, `aiCitedUrl`). `from` and `to` are optional ISO dates.\n\n### All tracked keywords\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n https://your-rankme-host/api/v1/keywords\n```\n\nReturns every tracked keyword across your sites with its latest position, delta, and AI Overview fields.\n\n## Compatibility\n\nThis release exposes exactly four read-only routes:\n\n- `GET /api/v1/sites`\n- `GET /api/v1/sites/:siteId/report/latest`\n- `GET /api/v1/sites/:siteId/rank-history`\n- `GET /api/v1/keywords`\n\nBrand Radar, Review Intelligence, Link Intelligence, Traffic Insights, Keyword Trends, and the SERP Sensor do not add routes under `/api/v1`. The five paid surfaces remain authenticated dashboard features; the public sensor remains under `/api/volatility`. Existing response fields keep their meaning, and clients should ignore additive fields they do not recognize.\n\n\n\n## Rate limits\n\nBy default each key may make **120 requests per minute**. Above that the API answers `429` until the window resets.\n\n## Errors\n\nErrors use the shape `{ \"error\": { \"message\": \"...\", \"details\": ... } }`, localized by your `Accept-Language` header:\n\n- `401` — the key is missing, malformed, revoked, or unknown.\n- `402` — your plan does not include the API.\n- `404` — the site or report does not exist on your account.\n- `429` — rate limit exceeded (body: `{ \"error\": \"...\" }`).\n\n[Back to the docs index](./index.en.md)"},{"slug":"rankmefast-mcp","locale":"en","title":"RankMeFast MCP","description":"Connect RankMeFast to coding agents, IDEs, and MCP-ready CLIs.","section":"developers","order":2,"searchText":"# Connect RankMeFast to your AI tools\n\nRankMeFast has a remote [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) endpoint. It lets a compatible coding agent, IDE, or CLI read your SEO data and start audits without leaving your current workflow.\n\n## Set up in five minutes\n\n1. Open [Settings → API keys](/profile?tab=api-keys), select **Create key**, and copy the key when it appears. RankMeFast shows the full key only once.\n2. Use the RankMeFast MCP endpoint:\n\n ```text\n https://rankme.fast/api/mcp\n ```\n\n3. For clients that support environment variables, set the key before starting the client:\n\n ```bash\n export RANKMEFAST_API_KEY='rmf_REPLACE_WITH_YOUR_KEY'\n ```\n\n In PowerShell:\n\n ```powershell\n $env:RANKMEFAST_API_KEY = 'rmf_REPLACE_WITH_YOUR_KEY'\n ```\n\n4. Choose your client below, paste its configuration, and restart or reload the client.\n5. Verify it with: **“Use RankMeFast to list my sites.”**\n\n> Treat the key like a password. RankMeFast API keys are account-wide and do not have separate scopes. The `start_audit` tool can consume your monthly audit allowance. Prefer an environment variable, password prompt, or user-level config, and never commit a key to your repository.\n\nMCP access is included with Starter, Pro, and Agency. Agency keys also work with the read-only public REST API.\n\n## Choose your client\n\n| Client or surface | Direct setup | Configuration |\n| --- | --- | --- |\n| Claude Code and its desktop Code tab | Yes | HTTP server with bearer header |\n| Cursor IDE and Cursor Agent CLI | Yes | User or project MCP JSON |\n| VS Code with GitHub Copilot | Yes | User or workspace `mcp.json` |\n| GitHub Copilot CLI | Yes | CLI command or user JSON |\n| Windsurf / Cascade | Yes | User MCP JSON |\n| Codex CLI, IDE extension, and ChatGPT desktop Codex | Yes | Shared Codex TOML |\n| Gemini CLI | Yes | CLI command or user JSON |\n| OpenCode | Yes | Remote MCP JSON |\n| JetBrains AI Assistant and Junie | Yes | IDE MCP settings |\n| Zed | Yes | Remote context server |\n| Cline | Yes | Streamable HTTP server |\n| Roo Code | Yes | Streamable HTTP server |\n| Kiro IDE and CLI | Yes | User or project MCP JSON |\n| Copilot in Visual Studio, JetBrains, Xcode, and Eclipse | Yes | Copilot MCP JSON |\n| Claude.ai / Claude Desktop chat connector | Not directly | The connector requires OAuth; RankMeFast currently uses bearer keys |\n| ChatGPT web | Not directly | It does not read local Codex MCP configuration |\n\nAny other client can connect if it supports remote **Streamable HTTP** and a custom `Authorization` header. RankMeFast does not expose a local stdio or legacy SSE server.\n\n## Claude Code\n\nAdd a user-level server. The single quotes keep the environment variable reference intact:\n\n```bash\nclaude mcp add-json --scope user rankmefast \\\n '{\"type\":\"http\",\"url\":\"https://rankme.fast/api/mcp\",\"headers\":{\"Authorization\":\"Bearer ${RANKMEFAST_API_KEY}\"}}'\n```\n\nCheck the connection with:\n\n```bash\nclaude mcp get rankmefast\n```\n\nYou can also run `/mcp` inside Claude Code. The Claude Code surface in the desktop app uses the same configuration. See the official [Claude Code MCP guide](https://code.claude.com/docs/en/mcp).\n\nThe Claude.ai and Claude Desktop **chat connector** is different: its remote connector flow documents OAuth, not an arbitrary static bearer header. It cannot connect directly until RankMeFast offers OAuth; use Claude Code instead. See [Claude custom connectors](https://support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp).\n\n## Cursor IDE and Cursor Agent CLI\n\nCreate `~/.cursor/mcp.json` for a user-level setup. Replace both placeholders; keep this file private:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n}\n```\n\nCursor IDE and Cursor Agent CLI read the same configuration. Verify it with:\n\n```bash\ncursor-agent mcp list\ncursor-agent mcp list-tools rankmefast\n```\n\nFor a team configuration, use `.cursor/mcp.json` but do not put a literal key in that committed file. See the official [Cursor MCP guide](https://docs.cursor.com/context/model-context-protocol).\n\n## VS Code and GitHub Copilot\n\nRun **MCP: Open User Configuration** from the Command Palette, then use a password input so the secret is not written into the file:\n\n```json\n{\n \"inputs\": [\n {\n \"type\": \"promptString\",\n \"id\": \"rankmefast-key\",\n \"description\": \"RankMeFast API key\",\n \"password\": true\n }\n ],\n \"servers\": {\n \"rankmefast\": {\n \"type\": \"http\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer ${input:rankmefast-key}\"\n }\n }\n }\n}\n```\n\nRun **MCP: List Servers** to start or inspect it. Workspace configuration can live in `.vscode/mcp.json`; the password-input form is safe to share. See the official [VS Code MCP configuration reference](https://code.visualstudio.com/docs/agents/reference/mcp-configuration).\n\n## GitHub Copilot CLI\n\nWith `RANKMEFAST_API_KEY` set in your shell:\n\n```bash\ncopilot mcp add \\\n rankmefast \\\n --type http \\\n --url https://rankme.fast/api/mcp \\\n --header \"Authorization=Bearer $RANKMEFAST_API_KEY\" \\\n --tools \"*\"\n```\n\nThe shell expands the key before Copilot saves the user configuration, so protect `~/.copilot/mcp-config.json`. See [Adding MCP servers to Copilot CLI](https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/add-mcp-servers).\n\n## Windsurf / Cascade\n\nAdd this to `~/.codeium/windsurf/mcp_config.json`:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"serverUrl\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer ${env:RANKMEFAST_API_KEY}\"\n }\n }\n }\n}\n```\n\nReload it from **Windsurf Settings → Cascade → MCP Servers**. See the official [Windsurf MCP guide](https://docs.windsurf.com/windsurf/cascade/mcp).\n\n## Codex CLI, IDE extension, and ChatGPT desktop\n\nCodex CLI, the Codex IDE extension, and the Codex surface in ChatGPT desktop share `~/.codex/config.toml` on the same machine:\n\n```toml\n[mcp_servers.rankmefast]\nurl = \"https://rankme.fast/api/mcp\"\nbearer_token_env_var = \"RANKMEFAST_API_KEY\"\n```\n\nRestart the IDE or desktop app after setting the environment variable. In the CLI, check it with:\n\n```bash\ncodex mcp list\n```\n\nYou can also run `/mcp` in a Codex session. See the official [Codex MCP guide](https://learn.chatgpt.com/docs/extend/mcp).\n\nChatGPT web does not read local Codex configuration, so this setup applies only to Codex surfaces on the configured machine.\n\n## Gemini CLI\n\nWith `RANKMEFAST_API_KEY` set in your shell:\n\n```bash\ngemini mcp add \\\n --scope user \\\n --transport http \\\n --header \"Authorization: Bearer $RANKMEFAST_API_KEY\" \\\n rankmefast https://rankme.fast/api/mcp\n```\n\nVerify it with `gemini mcp list` or `/mcp` inside Gemini CLI. The command stores the expanded header in `~/.gemini/settings.json`, so keep that file private. In JSON, use `httpUrl`; Gemini reserves `url` for legacy SSE. See the official [Gemini CLI MCP guide](https://github.com/google-gemini/gemini-cli/blob/main/docs/tools/mcp-server.md).\n\n## OpenCode\n\nAdd this to the global `~/.config/opencode/opencode.json` file:\n\n```json\n{\n \"$schema\": \"https://opencode.ai/config.json\",\n \"mcp\": {\n \"rankmefast\": {\n \"type\": \"remote\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"enabled\": true,\n \"oauth\": false,\n \"headers\": {\n \"Authorization\": \"Bearer {env:RANKMEFAST_API_KEY}\"\n }\n }\n }\n}\n```\n\n`oauth: false` prevents OAuth discovery because RankMeFast uses a bearer key. Check the setup with:\n\n```bash\nopencode mcp list\nopencode mcp debug rankmefast\n```\n\nSee the official [OpenCode MCP guide](https://opencode.ai/docs/mcp-servers/).\n\n## JetBrains AI Assistant and Junie\n\nOpen **Settings → Tools → AI Assistant → Model Context Protocol (MCP)**, add an HTTP server, and paste:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n}\n```\n\nKeep this in IDE-level settings rather than a project file. Enable **Pass custom MCP servers** when you want Junie or another integrated agent to receive the tools. See the official [JetBrains MCP guide](https://www.jetbrains.com/help/ai-assistant/mcp.html).\n\n## Zed\n\nOpen **Settings → AI → MCP Servers → Add Remote Server**, or add this to your user settings:\n\n```json\n{\n \"context_servers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n}\n```\n\nKeep the literal key in user settings, not project settings. A green indicator beside the server confirms the connection. See the official [Zed MCP guide](https://zed.dev/docs/ai/mcp).\n\n## Cline\n\nOpen Cline's MCP settings and add a Streamable HTTP server:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"type\": \"streamableHttp\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n },\n \"disabled\": false,\n \"autoApprove\": []\n }\n }\n}\n```\n\nLeave `autoApprove` empty so starting an audit requires your approval. See the official [Cline MCP guide](https://docs.cline.bot/mcp/mcp-overview).\n\n## Roo Code\n\nUse the global MCP settings or `.roo/mcp.json`:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"type\": \"streamable-http\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n },\n \"disabled\": false,\n \"alwaysAllow\": []\n }\n }\n}\n```\n\nLeave `alwaysAllow` empty so spending actions need approval. See the official [Roo Code MCP guide](https://docs.roocode.com/features/mcp/using-mcp-in-roo).\n\n## Kiro IDE and CLI\n\nUse `~/.kiro/settings/mcp.json` globally or `.kiro/settings/mcp.json` for a project:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer ${RANKMEFAST_API_KEY}\"\n }\n }\n }\n}\n```\n\nKiro expands environment variables in headers. See the official [Kiro MCP configuration](https://kiro.dev/docs/mcp/configuration/).\n\n## Copilot in other IDEs\n\nGitHub Copilot in Visual Studio, JetBrains IDEs, Xcode, and Eclipse uses a different shape from VS Code:\n\n```json\n{\n \"servers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"requestInit\": {\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n }\n}\n```\n\nStore this as user-level configuration and do not commit the key. Follow GitHub's [MCP setup guide for Copilot IDE extensions](https://docs.github.com/en/copilot/how-tos/provide-context/use-mcp-in-your-ide/extend-copilot-chat-with-mcp) for the config-file location in your IDE.\n\n## Available RankMeFast tools\n\n| Tool | What it does | Uses plan allowance? |\n| --- | --- | --- |\n| `list_sites` | Lists sites owned by your account | No |\n| `get_latest_audit_report` | Returns the latest completed audit for a site | No |\n| `list_keywords` | Lists tracked keywords for a site | No |\n| `get_rank_history` | Returns keyword rank history for a site | No |\n| `list_content_analyses` | Lists Content Intelligence analyses | No |\n| `get_content_analysis` | Returns one Content Intelligence analysis | No |\n| `start_audit` | Starts a site audit, within your plan's page cap | **Yes — one audit** |\n| `get_audit_status` | Checks an audit run | No |\n\nReads and actions are always scoped to the account that owns the key. A site, analysis, or run ID from another account is returned as not found.\n\nThe tool list changes additively: existing tools keep their names and argument shapes, and new tools may appear as features ship. Nothing in this release removes or renames a tool.\n\n## Troubleshooting\n\n- **The client cannot discover tools:** confirm the URL ends in `/api/mcp`, choose Streamable HTTP rather than SSE or stdio, then restart the client.\n- **401 Unauthorized:** use the exact `Authorization: Bearer rmf_…` header. The key may be mistyped, revoked, or attached to an inactive account.\n- **402 Upgrade required:** MCP needs Starter, Pro, or Agency. See [Plans, limits & credits](./plans-limits-credits.en.md).\n- **405 Method not allowed in a browser:** this is expected because the endpoint accepts MCP `POST` requests, not ordinary browser `GET` requests.\n- **429 Too many requests:** wait for the rate-limit window shown in the response headers, then retry.\n- **503 Unavailable:** MCP is disabled or temporarily unavailable on that RankMeFast installation; ask its operator.\n- **Works locally but not through a proxy:** make sure the proxy forwards `Authorization`, `Content-Type`, and `Accept` headers to `/api/mcp`.\n\nTo replace a key, create a new one, update and verify every client, then revoke the old key from [Settings → API keys](/profile?tab=api-keys). Revocation is immediate.\n\n## Compatibility\n\nThe registry in this release is exactly the eight tools listed above. Seven are read-only; `start_audit` is the only tool that spends a plan allowance. Brand Radar, Review Intelligence, Link Intelligence, Traffic Insights, Keyword Trends, and the SERP Sensor do not have MCP tools. No existing tool name or argument shape changed.\n\n\n\n## Related guides\n\n- [Public API](./public-api.en.md) — the Agency-only read API.\n- [Content Intelligence](./content-intelligence.en.md) — understand the analysis data returned by MCP.\n- [Account security](./settings-security.en.md) — keep your account and keys safe."}]}},"actionData":null,"errors":null}</article></body></html>

{"loaderData":{"0":null,"0-22":{"locale":"en","home":{"slug":"index","locale":"en","title":"RankMeFast documentation","description":"Guides for using RankMeFast, in plain language.","section":"start","order":0},"docs":[{"slug":"getting-started","locale":"en","title":"Getting started","description":"Sign up, add your site, run your first audit.","section":"start","order":1},{"slug":"audit-report","locale":"en","title":"Reading your audit report","description":"Fix now, Watch, Passed — and what every rule means.","section":"start","order":2},{"slug":"troubleshooting","locale":"en","title":"Troubleshooting","description":"Common problems and how to fix them.","section":"start","order":3},{"slug":"changelog","locale":"en","title":"Changelog","description":"Plain-language notes about product changes that affect how you use RankMeFast.","section":"start","order":4},{"slug":"page-speed","locale":"en","title":"Page speed","description":"Core Web Vitals in plain words — real visitors vs a lab estimate.","section":"audits","order":1},{"slug":"ai-summary","locale":"en","title":"AI summary","description":"What it does, what we send, and why it never changes the audit.","section":"audits","order":2},{"slug":"content-intelligence","locale":"en","title":"Content Intelligence","description":"Analyze a page you own, score it, and get a plain-language brief with citations.","section":"audits","order":3},{"slug":"ai-visibility","locale":"en","title":"AI Visibility","description":"Track whether AI assistants mention your brand. Agency feature.","section":"audits","order":4},{"slug":"local-seo","locale":"en","title":"Local SEO","description":"Business listing accuracy, reviews, and local map rank. Pro feature.","section":"audits","order":5},{"slug":"backlinks-competitors","locale":"en","title":"Backlinks & competitors","description":"Domain rating, link rows, competitor overlap.","section":"audits","order":6},{"slug":"weekly-pulse","locale":"en","title":"Weekly Pulse","description":"One deterministic weekly digest per site, from signals you have already paid for.","section":"audits","order":7},{"slug":"confirmed-rank-alerts","locale":"en","title":"Confirmed rank alerts","description":"How a rank change becomes confirmed, and what the observation labels mean.","section":"audits","order":8},{"slug":"next-actions","locale":"en","title":"Next Actions","description":"One prioritized list of what to do next, with a full decision history.","section":"audits","order":9},{"slug":"ai-visibility-citations","locale":"en","title":"AI citations & source gaps","description":"Sampled prompt checks — which sources AI answers cite, and where you are missing.","section":"audits","order":10},{"slug":"rank-tracking","locale":"en","title":"Rank tracking","description":"Adding keywords, reading positions, weekly vs daily.","section":"research","order":1},{"slug":"keyword-research","locale":"en","title":"Keyword research","description":"Volume and difficulty in plain words. Pro feature.","section":"research","order":2},{"slug":"google-search-console","locale":"en","title":"Connecting Google Search Console","description":"Why connect, what we access, how to disconnect.","section":"research","order":3},{"slug":"gsc-generative-appearance","locale":"en","title":"Generative AI appearance in Search Console","description":"A first-party GSC signal — how often your site was shown inside Google generative AI features.","section":"research","order":4},{"slug":"audience-research","locale":"en","title":"Audience Research","description":"Turn topics, competitors, and a market into audience segments from public evidence.","section":"research","order":5},{"slug":"keyword-intelligence","locale":"en","title":"Keyword Intelligence","description":"Gap, overview, trends, and clustering — what each check costs and what the labels mean.","section":"research","order":6},{"slug":"link-intelligence","locale":"en","title":"Link Intelligence","description":"Run deeper backlink checks and compare link gaps without guessing what a unit covers.","section":"research","order":7},{"slug":"traffic-insights","locale":"en","title":"Traffic Insights","description":"Compare modeled competitor traffic while keeping estimates, units, and refunds clear.","section":"research","order":8},{"slug":"keyword-trends","locale":"en","title":"Keyword Trends","description":"Explore search-interest direction, momentum, and seasonality with honest estimate labels.","section":"research","order":9},{"slug":"review-intelligence","locale":"en","title":"Review Intelligence","description":"Sync public reviews on demand, read trends and cited themes, and understand exactly when a unit is returned.","section":"research","order":10},{"slug":"brand-radar","locale":"en","title":"Brand Radar","description":"Scan one brand query on demand, inspect retained mentions, and understand the add-on and refund boundary.","section":"research","order":11},{"slug":"serp-sensor","locale":"en","title":"SERP Sensor methodology","description":"How the public daily volatility index is collected, calculated, budgeted, and limited.","section":"research","order":12},{"slug":"pricing","locale":"en","title":"Pricing","description":"How RankMeFast plans and monthly allowances work.","section":"account","order":1},{"slug":"plans-limits-credits","locale":"en","title":"Plans, limits & credits","description":"Starter, Pro, Agency — caps, trial, overage credits.","section":"account","order":2},{"slug":"settings-security","locale":"en","title":"Account security","description":"Change your password and manage account credentials.","section":"account","order":3},{"slug":"notification-preferences","locale":"en","title":"Notification preferences","description":"Choose which non-security emails you want to receive.","section":"account","order":4},{"slug":"team","locale":"en","title":"Team & invites","description":"Invite teammates, accept invites, and manage seats on your plan.","section":"account","order":5},{"slug":"public-api","locale":"en","title":"Public API","description":"Read your sites, reports, rankings, and keywords over a simple key-authenticated API. Agency feature.","section":"developers","order":1},{"slug":"rankmefast-mcp","locale":"en","title":"RankMeFast MCP","description":"Connect RankMeFast to coding agents, IDEs, and MCP-ready CLIs.","section":"developers","order":2}]},"0-22-0":{"locale":"en","docs":[{"slug":"getting-started","locale":"en","title":"Getting started","description":"Sign up, add your site, run your first audit.","section":"start","order":1,"searchText":"# Getting started\n\nWelcome. RankMeFast watches how well your site attracts visitors from Google — and, increasingly, from AI assistants. This guide takes you from sign-up to your first report.\n\n## 1. Create an account\nSign up with email or Google. If you use email, click the verification link we send. It expires after 24 hours; check spam if it does not arrive in two minutes.\n\n## 2. Add your site\nEnter the full URL including `https://`. We only accept public URLs that respond, and we normalise to the scheme + host (paths and query strings are dropped). Local addresses like `localhost` or IP addresses are rejected.\n\n## 3. Run your first audit\nClick **Run audit**. The audit takes 1–3 minutes: we crawl a small set of your pages, check technical signals, and score field/lab performance. Free accounts get one full audit at sign-up; paid accounts get several per month (see [Plans, limits & credits](./plans-limits-credits.en.md)).\n\n## 4. Read the three tabs\n- **Fix now** — the problems Google reacts to fastest. Do these first.\n- **Watch** — smaller issues or checks where we do not have enough data yet.\n- **Passed** — everything your site cleared.\n\nEach row explains **why** the problem matters, **which page** it was found on, and **the exact fix**. Full walkthrough: [Reading your audit report](./audit-report.en.md).\n\n## 5. Optional but recommended\n- Connect Google Search Console so RankMeFast can tell whether your pages are indexed and whether rich results appear. See [Connecting Google Search Console](./google-search-console.en.md).\n- Add a handful of keywords to track (in the site view) so you can watch positions over time. See [Rank tracking](./rank-tracking.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"audit-report","locale":"en","title":"Reading your audit report","description":"Fix now, Watch, Passed — and what every rule means.","section":"start","order":2,"searchText":"# Reading your audit report\n\nEvery audit lands you on a screen with three tabs and a Retest button. Everything here is written for a non-technical site owner — no SEO jargon without its plain meaning.\n\n## The three buckets\n- **Fix now** — problems Google reacts to fastest. Every row is either critical or a broken signal you can act on today.\n- **Watch** — smaller issues, or checks where we did not have enough data to score. Worth tracking, not usually worth stopping the world.\n- **Passed** — checks your site cleared. Reviewing this tab occasionally is a good sanity check.\n\n## Anatomy of an issue\nEach row shows:\n- **Title** — a plain-language name for the problem.\n- **Why it matters** — one sentence on why Google or a real visitor cares.\n- **Where** — the exact URL(s) affected, so you can jump straight there.\n- **Exact fix** — the concrete change to make. A copy button puts the fix on your clipboard.\n\n## Retest and diff badges\nClick **Retest** after you have shipped a fix. The next audit compares against the last one and shows two badges on each row:\n- **Fixed** — the rule that used to fail now passes on that URL.\n- **Regressed** — the rule newly failed on a URL. New failures also count as regressions.\n\nDiffs are taken against the frozen snapshot of your previous audit, so what you saw yesterday will still say the same thing tomorrow.\n\n## `llms.txt`\nA `/llms.txt` file is an emerging convention: a plain-text pointer file that lets AI assistants (ChatGPT, Perplexity, Claude, and others) know which of your pages are safe to cite and preferred. RankMeFast treats this rule as **advisory** — we suggest adding one, but nothing is broken if you have not.\n\n## Every rule in plain words\n- **Blocked from Google** — A `robots.txt` rule is telling Google to stay away. If the page should rank, remove the disallow line.\n- **Sitemap missing or weak** — A sitemap is a file at `/sitemap.xml` that lists your pages. Add one and reference it from `robots.txt`.\n- **Page titles missing or weak** — The `` element decides your Google click bait. Aim for a specific, 30–60 character title on every page.\n- **Meta descriptions missing or duplicated** — The one-line snippet Google shows below the title. Write a unique description per page.\n- **Page headings missing or overused** — One `<h1>` per page, then `<h2>` / `<h3>` for structure. Missing `<h1>` is a big deal; two `<h1>`s is a smaller one.\n- **Canonical link missing or broken** — A `<link rel=\"canonical\">` tells Google which URL is the \"real\" one. Missing or pointing off-site is bad.\n- **Structured data missing or broken** — Machine-readable tags (JSON-LD) that help Google understand articles, products, FAQs.\n- **Broken internal links** — Links from one of your pages to another that return an error. Fix or remove them.\n- **Thin content** — Pages with fewer than ~200 words look empty to Google. Expand the ones that should rank.\n- **No FAQ signals found** — No question-and-answer sections detected. FAQs help both users and AI assistants.\n- **llms.txt file missing (advisory)** — An emerging convention: a `/llms.txt` file that hints AI crawlers about your best content. Not required — advisory only.\n- **HTTPS not enforced** — Your site should redirect http → https and canonicalise the https version.\n- **Poor real-visitor performance** — Real visitors experienced a slow page. Data comes from Google’s Chrome UX Report — sites with little traffic may have none of it yet.\n- **Low lab performance estimate** — Our synthetic Lighthouse run scored low. A single run varies ~10 points; treat as a hint, not a verdict.\n- **Not mobile-friendly** — Phones do not render the page well. Check tap-target sizes and the viewport meta tag.\n- **Not indexed by Google** — Google looked at the page but did not add it to search. Reconnect Google Search Console if this rule cannot fetch data.\n- **Rich results issues** — Structured data was recognised, but Google flagged errors that block enhanced search results.\n- **Indexed with warnings** — The page is indexed but Google flagged a canonicalization or duplicate-content signal.\n\n[Back to the docs index](./index.en.md)"},{"slug":"troubleshooting","locale":"en","title":"Troubleshooting","description":"Common problems and how to fix them.","section":"start","order":3,"searchText":"# Troubleshooting\n\n## Audit failed or unavailable\nAn audit fails when we cannot reach your site or the crawl provider timed out. Common causes:\n- Your site is behind a Cloudflare \"under attack\" rule blocking crawlers.\n- Robots.txt blocks user-agent `RankMeFastBot` or \"`*`\".\n- Origin returned 5xx for our first requests.\n\nTry again in a few minutes. If it keeps failing, adjust firewall/robots.txt to allow our bot.\n\n## Rank check unavailable\nA single keyword can come back unavailable if the SERP provider is rate-limited or Google served a captcha. It is temporary — the next scheduled run retries. Nothing to do.\n\n## \"Reconnect needed\" on Search Console\nGoogle invalidated our token. Open Settings → Google Search Console → **Reconnect**. This can happen if you changed your Google password or revoked the app.\n\n## Cap reached\nSee [Plans, limits & credits](./plans-limits-credits.en.md). Options are: upgrade tier, buy a small credit pack, or wait until the next billing period.\n\n## Cap reached on a research surface (402)\nThe response names the exact unit that ran out — for example `keyword_lookups` or `audience_research_runs`. Wait for the monthly reset, buy a credit pack where one exists, or upgrade; Audience Research runs have no pack.\n\n## Partial results on a research surface\nSome engines or sources answered, others did not. The result ships with what arrived and is labeled **partial results** — the missing part is named, never zero-filled.\n\n## Unsupported for this engine / market\nThe check cannot run for that engine or market. The unsupported part is excluded from the result — it is never displayed as a zero.\n\n## Unavailable (not zero)\nNo data came back for the window. `unavailable ≠ 0`: treat it as unknown, not as a drop to zero. You will see this on share-of-voice and Search Console generative-AI cards.\n\n## \"Reconnect needed\" on analytics or generative-AI cards\nSame cause as the Search Console reconnect above — Google invalidated the token. Reconnect from the site's Google workspace; the cards recover on the next read.\n\n## Verification email missing\nCheck spam. The link expires after 24 hours. If it truly did not arrive:\n- Confirm you typed the address correctly (typos are the #1 cause).\n- Try Google sign-in instead if your address is Gmail.\n- Contact support if the address is a corporate one — the mail may be filtered upstream.\n\n## AI summary shows \"shortened\"\nLong audits are truncated before we send them to the model — the summary is still valid, just from a shortened view. Click **Regenerate** if you want another take.\n\n## Report screen loops loading\nHard-refresh the page (⌘/Ctrl + Shift + R). Clear the local cache if that does not help. If the issue persists, it usually points at a mismatch between an in-flight audit and the report snapshot — support can force-refresh from their side.\n\n[Back to the docs index](./index.en.md)"},{"slug":"changelog","locale":"en","title":"Changelog","description":"Plain-language notes about product changes that affect how you use RankMeFast.","section":"start","order":4,"searchText":"# Changelog\n\n## 29 July 2026 — New intelligence tools, no price change for existing subscriptions\n\nWe added [Link Intelligence](./link-intelligence.en.md), [Traffic Insights](./traffic-insights.en.md), [Keyword Trends](./keyword-trends.en.md), [Review Intelligence](./review-intelligence.en.md), [Brand Radar](./brand-radar.en.md), and the public [SERP Sensor](./serp-sensor.en.md).\n\nIf you already subscribe, your current RankMeFast subscription price is unchanged. The new monthly allowances, optional Brand Radar add-on, and one-time packs apply as described on [Pricing](./pricing.en.md). Add-on and pack prices apply only if you choose to buy them.\n\nWe are not moving you to another plan, and you do not need to take action. Your existing stored reports and history remain available under the rules of your current plan.\n\n[Back to the docs index](./index.en.md)"},{"slug":"page-speed","locale":"en","title":"Page speed","description":"Core Web Vitals in plain words — real visitors vs a lab estimate.","section":"audits","order":1,"searchText":"# Page speed\n\nGoogle grades every page on **Core Web Vitals** — three signals about how it feels to load and interact with your page. There are two flavours of measurement and RankMeFast shows both:\n\n## Real-visitor data (field data)\nGoogle’s Chrome UX Report aggregates real visits from Chrome users. When your site has enough traffic, this is the number that matters — it is what Google will judge you on. The three signals:\n- **LCP (Largest Contentful Paint)** — how quickly the biggest visible element appears. Aim under 2.5 seconds.\n- **INP (Interaction to Next Paint)** — how snappy your page feels when you click or tap. Aim under 200 ms.\n- **CLS (Cumulative Layout Shift)** — how much things jump around during load. Aim under 0.1.\n\nIf your site is new or has low traffic, field data may not exist yet — the row shows \"not enough visitor data yet\". That is normal.\n\n## Lab estimate (Lighthouse)\nWhen there is no field data, RankMeFast still runs a synthetic Lighthouse test in a controlled environment. Treat this as a **hint**: a single Lighthouse run varies by ~10 points from one attempt to the next. Do not chase small changes — look at the trend across audits.\n\n## Mobile\nMobile is measured separately because phones render pages very differently. If your page is not mobile-friendly (missing viewport meta tag, tap targets too small), it lands in **Fix now**.\n\nTip: page-speed problems are usually image-shaped. Compress big images, serve modern formats (`webp` / `avif`), and lazy-load anything below the fold.\n\n[Back to the docs index](./index.en.md)"},{"slug":"ai-summary","locale":"en","title":"AI summary","description":"What it does, what we send, and why it never changes the audit.","section":"audits","order":2,"searchText":"# AI summary\n\nThe AI summary sits above the tabs on the report screen. It reads your audit and writes a short paragraph in your language telling you which fixes to do first.\n\n## What it does\nThe summary is a plain-language executive brief: which findings are most important, why they matter, and roughly which one to start with. It does not add new findings — it re-tells the ones the audit already produced.\n\n## What we send\nOnly two things are sent to the AI model:\n1. The text of your audit findings — the same rule titles, plain-word explanations, and fix strings you can see on the screen.\n2. Your affected URLs, so it can reference them.\n\nWe do **not** send Search Console data, screenshots, your billing info, or anything about other sites on your account.\n\n## Why it never affects the audit itself\nThe audit is deterministic — same input, same output. The AI summary is a separate step that runs *after* the audit finishes, using the finished snapshot as read-only input. Turning the summary off has no effect on any rule’s bucket.\n\n## Regenerating\nClick **Regenerate** on the summary card to run it again. Each regeneration uses one summary credit (see the AI summaries cap in [Plans, limits & credits](./plans-limits-credits.en.md)). If you hit the cap the card shows a paywall notice; unpaid retries do not spend credit.\n\n## Choosing a model\nThe summary uses a small, low-cost Claude model by default. Some findings — very long lists — are shortened before sending. If shortening happens, the card shows a \"shortened\" note.\n\n[Back to the docs index](./index.en.md)"},{"slug":"content-intelligence","locale":"en","title":"Content Intelligence","description":"Analyze a page you own, score it, and get a plain-language brief with citations.","section":"audits","order":3,"searchText":"# Content Intelligence\n\nContent Intelligence looks at a page you own and returns a deterministic\nreadiness score, the evidence behind it, a plain-language brief, and — if\nyou opt in — an AI-generated draft. One analysis draws one\n`content_analyses` credit from your monthly allowance or your credit pack.\n\n## What it does\n\n- Fetches the page through the Firecrawl provider you own the crawl for.\n- Extracts headings, links, structured data, and a small set of on-page\n signals. No third-party HTML is stored raw.\n- Scores the page against the same rule engine that powers the audit report,\n so the numbers stay comparable.\n- Returns evidence for every finding: the rule id, where it fired, the\n citation excerpt, and a confidence label.\n- If you opt in, calls the AI provider order from your operator config to\n produce a brief and (optionally) a draft. The provider selection is\n server-side; you never see raw provider payloads.\n\n## What it does NOT do\n\n- It never publishes anything. Drafts are yours to accept, edit, or dismiss.\n- It never crawls arbitrary URLs. The URL must belong to a site you own or\n a competitor you have added on the Agency tier.\n- It never returns the raw provider payload, prompt, or completion. The\n public API response is the same normalized shape the app renders.\n\n## Starting an analysis\n\nYou can start an analysis from:\n\n- **Sites → any site → Content** — pick a page URL from your inventory.\n- **Report → Fix now / Watch** — the recommendation row opens the workflow\n pre-filled.\n- **Keyword research → tracked keyword** — analyzes the URL that ranks for\n the keyword right now.\n- **Google Search Console → any query** — analyzes the URL Google surfaces\n for that query.\n- **Competitors → Agency competitor page** — Agency tier only.\n\nEach entry point deep-links into the same workflow. Every one of them\nreserves capacity before enqueueing (see\n[Plans, limits & credits](./plans-limits-credits.en.md)); the flow refuses\nloudly if you are over your cap.\n\n## What one credit includes\n\nOne `content_analyses` credit covers:\n\n- one page fetch through Firecrawl,\n- the deterministic score and evidence,\n- the plain-language brief,\n- and — when opt-in AI is enabled — one draft.\n\nRegeneration draws another credit. Retrying the same URL within the\nseven-day snapshot window reuses the cached page but still costs one\ncredit for the new score and brief.\n\n## Score, evidence, confidence, citations, brief, draft\n\n- **Score.** Deterministic 0–100. Same rule engine as the audit report.\n- **Evidence.** Each finding names the rule, points at the exact HTML\n location it fired on, and shows a short excerpt.\n- **Confidence.** `high` / `medium` / `low`. Confidence is a function of\n page structure quality — it is not vendor spin.\n- **Citations.** Every AI claim cites the URL and excerpt it came from.\n Uncited claims are not shown.\n- **Brief.** A short, plain-language plan you can share with a writer.\n- **Draft.** Opt-in. The AI provider order in operator config decides who\n writes it. You accept, edit, or dismiss it — it is never auto-published.\n\n## Opt-in AI processing and retention\n\nThe AI brief and draft are opt-in. When you opt in, we send only the\nexcerpts needed to produce the brief. Excerpts are retained for seven days\nso we can regenerate on request without a second fetch, then purged. Your\nexport and account-deletion controls in\n[Account security](./settings-security.en.md) still apply and complete\ninside the documented grace window.\n\n## Partial and refunded results\n\nSometimes Firecrawl or the AI provider degrades. The analysis surfaces\nwhat completed successfully with a `partial` flag; you can regenerate the\nmissing sections at no extra credit. If the run failed hard before any\nuseful signal, the credit is refunded automatically and the analysis is\nmarked `refunded`.\n\n## Recommendations and 28-day correlations\n\nThe brief lists recommendations. You can:\n\n- **Accept** — moves the recommendation into your outcome loop.\n- **Dismiss** — records the dismissal reason for the correlation window.\n- **Apply** — marks it applied; the outcome loop measures the 28-day\n before/after correlation on rank, clicks, and impressions.\n\nThe correlation is descriptive, not causal — it is the same window we use\nin the audit report, and the number tells you whether the change moved.\n\n## Pro inventory and cannibalization\n\nOn Pro and Agency, the workspace opens a page inventory that groups URLs\nby topic and flags cannibalization pairs (two of your own URLs competing\non the same intent). Inventory scans do not draw analysis credits — they\nrun on the shared vendor cache.\n\n## Agency competitor portfolio and monitoring\n\nAgency adds a competitor portfolio: you pick competitor URLs, run analyses\nagainst them, and — with Content Monitoring on — get notified when a\ntracked competitor page changes materially. Monitoring uses the Firecrawl\nchange signature; a dismissed signature does not re-fire.\n\n## Managing allowances and the 20-analysis pack\n\nYour monthly `content_analyses` allowance depends on your tier — see\n[Plans, limits & credits](./plans-limits-credits.en.md). If you run out\nmid-month, the 20-analysis credit pack adds twenty single-use credits that\nnever expire and stack on top of the monthly quota.\n\n## Troubleshooting\n\n- **Access.** The URL must belong to a site you own (or an Agency\n competitor). Otherwise the flow returns \"not found\" — never \"forbidden\"\n — so it does not leak whether the URL exists.\n- **robots.txt.** If the page blocks crawlers, the analysis returns\n `partial` and cites the block; we never bypass robots.\n- **Stale GSC.** If Google has not yet indexed a URL, the analysis still\n runs on the fetched HTML but does not correlate against zero GSC data.\n- **Caps.** The flow refuses with a localized message that names the\n metric (`content_analyses`) — buy a pack or wait for the reset.\n- **Provider degradation.** Firecrawl or an AI provider outage surfaces\n as `partial` or `refunded` — never as a silent success.\n\n## Connecting the MCP surface\n\nIf you use Claude Code, Claude Desktop, Cursor, or VS Code, you can drive\nContent Intelligence from the RankMeFast MCP endpoint — see\n[RankMeFast MCP](./rankmefast-mcp.en.md).\n\n## Export and deletion\n\nEvery analysis is included in the account export. Deleting your account\npurges analyses inside the grace window documented in\n[Account security](./settings-security.en.md). We never publish anything\nyou have not explicitly accepted."},{"slug":"ai-visibility","locale":"en","title":"AI Visibility","description":"Track whether AI assistants mention your brand. Agency feature.","section":"audits","order":4,"searchText":"# AI Visibility\n\nAI assistants — the chat-style answers people increasingly get instead of a page of blue links — now decide what many buyers see first. AI Visibility tracks whether those answers mention your brand, point people to your site, or send them to a competitor. It is an **Agency** feature.\n\n## Tracked prompts\n\nA prompt is a question a buyer might ask an AI assistant — for example, \"best project management tool for a small team\". You add the prompts that matter to your business. You can track up to **10 prompts per site**. Pick the ones closest to a buying decision; that is where being mentioned is worth the most.\n\n## Running a check\n\nA check runs each tracked prompt and reads the answer the assistant gives back. For every prompt it records:\n\n- whether your brand is **mentioned** by name,\n- whether your site is **linked** as a source,\n- which **competitors** are named,\n- whether the mention is **positive, neutral, or negative**.\n\n\"Mentioned\" simply means your brand name appears in the answer. Being linked as a source is tracked separately and is the stronger signal.\n\n## Share of voice\n\nAcross all your tracked prompts, share of voice is the share of answers that mention you at all. Rising share of voice means you are showing up in more of the answers buyers see; a competitor climbing while you stay flat is an early warning.\n\n## An honest note\n\nThis is **opportunity tracking, not a ranking guarantee**. AI answers are generated fresh each time and can change between runs, between people, and between regions — the same prompt can mention you today and not tomorrow. Treat the trend across many checks as the signal, never a single result. Nothing here lets us, or anyone, promise a place in an AI answer.\n\n## Limits\n\nTracked prompts are capped at 10 per site, and each account has a monthly allowance of checks. When the allowance runs out, checks pause until the next monthly reset — your existing history stays visible. See [Plans, limits & credits](./plans-limits-credits.en.md) for the numbers.\n\n[Back to the docs index](./index.en.md)"},{"slug":"local-seo","locale":"en","title":"Local SEO","description":"Business listing accuracy, reviews, and local map rank. Pro feature.","section":"audits","order":5,"searchText":"# Local SEO\n\nIf customers find you by searching for a place — \"plumber near me\", \"coffee shop downtown\" — most of that traffic comes from the map results and business listings, not the regular blue links. Local SEO tracks the three things that decide whether you show up there. It is available on **Pro** and **Agency**.\n\n## Listing accuracy\n\nSearch engines trust a business whose details match everywhere they appear. RankMeFast checks that your **business name, address, and phone number** are consistent across the listings we can see. Mismatches — an old address, a second phone number, a shortened name — quietly push you down in local results. We flag each inconsistency so you can fix it at the source.\n\n## Reviews and questions\n\nWe track how many **reviews** you have, your **average rating**, and how fresh the reviews are, plus the public **questions and answers** on your listing. Steady, recent, well-answered reviews are one of the strongest local signals; a stall or a run of low ratings shows up here early.\n\n## Local map rank\n\nFor the keywords you already track, we also record where you appear in the **local map results** — the small pack of businesses shown above the regular results. This is separate from your normal search position: a site can rank well in ordinary results and still be missing from the map.\n\n## What one check costs\n\nOne **local check** refreshes your listing details, reviews, and questions together, counted as a single unit against your plan. Local map-rank checks are not separate — they use your normal rank-check budget. When your monthly local checks run out, they pause until the next reset, or you can top up with credits. See [Plans, limits & credits](./plans-limits-credits.en.md) for the exact numbers.\n\n[Back to the docs index](./index.en.md)"},{"slug":"backlinks-competitors","locale":"en","title":"Backlinks & competitors","description":"Domain rating, link rows, competitor overlap.","section":"audits","order":6,"searchText":"# Backlinks & competitors\n\nBacklinks are links from other sites to yours. Google treats them as a vote of confidence — the stronger the linking sites, the stronger your own site looks. Competitor analysis compares your link profile to other sites in your space.\n\n## Domain rating\nDomain rating is a 0–100 score of how strong a site’s **link profile** looks: the count and authority of sites linking to it. **Not** a measure of traffic or content quality. Ballpark:\n- 0–20 — small or new\n- 20–50 — established\n- 50–80 — strong\n- 80–100 — very few sites reach this\n\nTwo different providers score the same site slightly differently. Compare within the same tool over time, not across tools.\n\n## Row limits\nThe backlink list can be enormous — sites with a lot of history can have millions of rows. Each plan caps how many rows we pull per month:\n- Pro — up to a monthly cap on total rows.\n- Agency — unlimited rows.\n- Starter — no backlink data.\n\nSee [Plans, limits & credits](./plans-limits-credits.en.md) for the exact numbers.\n\n## Competitor overlap and the gap view\nAdd competitor domains and RankMeFast shows the **overlap** (keywords/backlinks you both target) and the **gap** (keywords they rank for that you do not). This is the fastest way to find low-hanging content ideas. Available on **Agency**.\n\nNote: numbers reflect what the provider has crawled. Recently launched or heavily paywalled sites can look weaker than they are.\n\n[Back to the docs index](./index.en.md)"},{"slug":"weekly-pulse","locale":"en","title":"Weekly Pulse","description":"One deterministic weekly digest per site, from signals you have already paid for.","section":"audits","order":7,"searchText":"# Weekly Pulse\n\nWeekly Pulse is an opt-in email digest that summarizes what changed for your site over the last ISO week. It is off by default. Each verified team member turns it on for themselves — nobody enables it for you.\n\n## What one pulse costs\n\nEmitting a digest for a site consumes **exactly one `ai_mentions_checks` unit** for that site, no matter how many teammates are subscribed. One collection, many recipients. Reading history, re-opening a past digest, or exporting stored data costs nothing and calls no provider.\n\n## What it contains\n\nThe digest is deterministic — it is rendered from signals we already stored:\n\n- New and lost citations across the AI engines your plan covers.\n- Confirmed rank drops from your tracked keywords.\n- Actions completed or regressed since the prior compatible pulse.\n- Your next three open actions.\n- Google Search Console generative-AI appearance, when Google returned rows.\n\nThe pulse never runs a fresh AI answer, and never issues a new SERP request.\n\n## When it runs\n\nEach site has a fixed weekly slot, deterministically derived from the site ID and spread across Monday–Saturday, 09:00–14:00 UTC. Sunday is reserved. You can see the next slot on the AI Visibility workspace.\n\n## States you may see\n\n- **Completed** — every supported engine returned data and the digest is ready.\n- **Partial** — some engines returned partial data; the digest still ships with what we have.\n- **Unsupported** — none of your plan's engines support this site's market and cohort; no unit is consumed.\n- **Blocked** — your `ai_mentions_checks` allowance is exhausted for the month; no provider call is made.\n- **Failed** — every supported engine failed for this run. The unit is consumed once and never refunded.\n\n## Turning it on or off\n\nOpen the site's AI Visibility workspace, review the spend preview, then toggle **Weekly Pulse**. Disabling suppresses future emails immediately. Re-enabling picks up the next weekly slot.\n\n## Privacy\n\nEmails contain bounded summaries only — host, count, keyword text, rank number, action verb, appearance name. Raw AI answers, source excerpts, competitor prose, prompts, and vendor task IDs never appear in the digest.\n\n[Back to the docs index](./index.en.md)"},{"slug":"confirmed-rank-alerts","locale":"en","title":"Confirmed rank alerts","description":"How a rank change becomes confirmed, and what the observation labels mean.","section":"audits","order":8,"searchText":"# Confirmed rank alerts\n\nA drop only becomes an alert after we observe it twice. That keeps noisy one-off movements out of your inbox.\n\n## How a change becomes confirmed\n\n1. A scheduled rank check finds that one of your top keywords fell by several positions.\n2. We run one fresh second observation for the same keyword in the same market — never a cached result.\n3. The second observation still shows the drop? Then it is a **confirmed rank change**, and only then do we alert you.\n\n## Volatile and unconfirmed changes\n\n- **Volatile** — the second observation shows the keyword recovered. We keep the record as history and send nothing.\n- **Unconfirmed** — we could not complete the second observation (monthly allowance exhausted, or the provider failed). It stays informational and never alerts.\n\nAll three outcomes remain visible next to the keyword's rank history, so nothing is hidden.\n\n## Observation labels\n\nEvery entry states exactly what we observed and when:\n\n- **Market** — the country, language, and device the observation used.\n- **Window** — the time of the first and the second observation.\n- **Freshness of the observation** — how current the data is. A stale observation is labeled as stale, never presented as real-time.\n- **Source** — which provider surface produced the observation.\n\n## Where alerts go\n\nA confirmed rank change appears in [Next Actions](./next-actions.en.md) with a direct link to the keyword's drop record. Dismiss or reopen it there — every decision is kept in the history.\n\n## Reading is free\n\nOpening the alert list, a drop record, or the history reads stored data only. A read never triggers a fresh SERP request and never consumes a unit from your plan.\n\n[Back to the docs index](./index.en.md)"},{"slug":"next-actions","locale":"en","title":"Next Actions","description":"One prioritized list of what to do next, with a full decision history.","section":"audits","order":9,"searchText":"# Next Actions\n\nNext Actions gathers everything worth doing into one prioritized list per site. No hunting through separate reports.\n\n## Where items come from\n\n- Audit findings from your latest finished audit (Fix now and Watch).\n- Confirmed rank changes from your tracked keywords.\n- Search Console and analytics declines.\n- Content recommendations and citation gaps from Content Intelligence.\n- Accepted signals from Audience Research.\n\n## How the list is ordered\n\nThe order is deterministic — the same evidence always produces the same list, and there is no hidden score. We sort by severity, then first-party impact, then confidence (high / medium / low), then how many pages are affected, then effort, then the newest observation; ties break on a stable identifier so the order never jumps.\n\n## Deciding on an item\n\nFour verbs: **Plan**, **Dismiss**, **Complete**, **Reopen**. Every decision is appended to the item's history — nothing is overwritten or deleted.\n\n## Dismissing a finding never edits the audit\n\nDismiss an audit finding as a false positive and it leaves the list. The audit report itself stays untouched — the finding remains in the report exactly as observed.\n\n## Retest\n\nRetest exists for audit findings only. It shows a spend preview first; running it uses one audit from your monthly allowance, the same as starting an audit yourself.\n\n## Reads are free\n\nOpening the list or an item's history reads stored data only. It never calls a provider and never consumes a unit.\n\n[Back to the docs index](./index.en.md)"},{"slug":"ai-visibility-citations","locale":"en","title":"AI citations & source gaps","description":"Sampled prompt checks — which sources AI answers cite, and where you are missing.","section":"audits","order":10,"searchText":"# AI citations & source gaps\n\nAI Visibility checks what AI assistants answer for the prompts you track, and which sources those answers cite. This page explains citations, source gaps, and the states around them.\n\n## A sample, never full coverage\n\nYou track a small set of prompts per site (up to ten). One check asks those prompts across the engines your plan covers, so the result is a sample — sample size (prompts × cohort) — of what people might see. It is never full coverage of everything people ask an AI.\n\n## Supported engines and surfaces\n\nMention checks read the provider's mentions index for Google and ChatGPT surfaces. Answer checks cover ChatGPT, Gemini, and Claude; Perplexity supports live answers only. New surfaces reported by the provider pass through under their own name — we never relabel them.\n\n## Citations\n\nA citation is a URL plus the identity of the source the answer pointed at. We record the exact cited URL whenever the engine reports one, and whether an answer cited you or a tracked competitor.\n\n## Source gaps\n\nA source gap means answers in your topic cite sources — but never yours. Source gaps surface as items in [Next Actions](./next-actions.en.md), each pointing at the sources the engines preferred.\n\n## States\n\n- **Partial results** — some engines returned data, others did not. We show what arrived and name what is missing.\n- **Unsupported for this engine / market** — the engine cannot run this check where you are. It is excluded, never shown as a zero.\n- **Unavailable (not zero)** — no data came back. `unavailable ≠ 0`: it means we do not know, not that you were never cited.\n\n## Not AI-market share\n\nShare of voice compares how often answers name you versus the competitors you track — inside your sampled prompts only. It is not a market-share measurement, and we never present it as one.\n\n## What a check costs\n\nOne check uses one `ai_mentions_checks` unit per tracked prompt. Reading stored results, citations, and source gaps is free.\n\n[Back to the docs index](./index.en.md)"},{"slug":"rank-tracking","locale":"en","title":"Rank tracking","description":"Adding keywords, reading positions, weekly vs daily.","section":"research","order":1,"searchText":"# Rank tracking\n\nRank tracking watches where your pages appear on Google for the keywords that matter to you.\n\n## Adding keywords\nOpen the site view and use **Add keyword**. Type the phrase you want to track. RankMeFast checks it against your keyword cap (see [Plans, limits & credits](./plans-limits-credits.en.md)) and refuses duplicates.\n\n## Reading positions\n- **Position** — the rank of your best-ranking page for that keyword. Position 1 means the top blue link. \"Not in top 100\" means we searched but did not find you in the first hundred results.\n- **Delta** — the change since last check. **-3** means you moved up three positions; **+5** means you slipped five.\n- **Last checked** — how recently the position was measured. Weekly checks land once a week; daily checks land every day.\n\n## Weekly vs daily\nEvery plan gets **weekly** checks bundled. **Daily** checks are a paid add-on — they are more expensive to run because they hit Google’s SERP much more often.\n\n## Trend chart\nSelecting a keyword shows a trend of your position over time. A rising line is bad (higher number = worse rank); a falling line is good. Under the chart is an accessible data table with the same numbers.\n\n## \"Unavailable\" checks\nSometimes a check comes back as unavailable. The most common reasons:\n- Our SERP provider’s quota is exhausted for the day.\n- The keyword returns a SERP feature we do not measure (news carousel only, for instance).\n- Google served a captcha and we could not verify the result.\nUnavailable checks do not consume credit and are retried on the next scheduled run.\n\n## Cap reached\nEvery plan has a keyword cap. When you hit it, the Add form is replaced with a notice pointing you to upgrade. Existing keywords keep running.\n\n[Back to the docs index](./index.en.md)"},{"slug":"keyword-research","locale":"en","title":"Keyword research","description":"Volume and difficulty in plain words. Pro feature.","section":"research","order":2,"searchText":"# Keyword research\n\nKeyword research helps you decide which phrases to target. It is a **Pro** and **Agency** feature — Starter accounts can upgrade to unlock it (see [Plans, limits & credits](./plans-limits-credits.en.md)).\n\n## Volume\nSearch volume is an estimate of how often a phrase is searched for on Google, per month, in the country you pick. It is an average smoothed over recent months — not a live counter. Values under 20 are noisy; treat them as directional.\n\n## Difficulty\nDifficulty is a 0–100 score of how hard it looks to rank on page 1 for the phrase. Higher = harder. The score blends how strong the current top-10 sites look and how many quality backlinks they have. A difficulty of 25 on a well-established site is easy; the same 25 on a brand-new domain is not.\n\n## Caching\nVolume and difficulty do not change hour-to-hour — so RankMeFast caches the answer for a while. If two people on your account look up the same phrase in the same country in the same week, the second request comes back instantly and does not count against your metered usage. If you re-run after the cache expires, we go back to the provider.\n\n## Suggestions and questions\nWhen available, we also show related phrases and question variants. Both use the same volume/difficulty numbers.\n\nNote: numbers come from a third-party provider and are rounded. Two different tools rarely agree exactly. Trust the ranking of one number against another more than the absolute value.\n\n[Back to the docs index](./index.en.md)"},{"slug":"google-search-console","locale":"en","title":"Connecting Google Search Console","description":"Why connect, what we access, how to disconnect.","section":"research","order":3,"searchText":"# Connecting Google Search Console\n\nSearch Console is Google’s free tool that shows how Google sees your site. Connecting it to RankMeFast lets three audit rules use the real answer instead of guessing:\n- **Not indexed by Google** — is a specific page actually in search?\n- **Rich results issues** — are your structured-data snippets valid?\n- **Indexed with warnings** — has Google flagged a canonicalization or duplicate signal?\n\n## Why connect\nWithout a connection, these rules stay in **Watch** with \"insufficient data\". With a connection, they land in the right bucket with actionable pages.\n\n## What we access (and what we do not)\n- We ask Google for the **read-only** Search Console scope: `webmasters.readonly`. That is the smallest scope that lets us call the URL Inspection API.\n- We **never** get access to Gmail, Drive, or your Google account beyond Search Console.\n- We store an encrypted refresh token so we can renew the short-lived access token. Nothing else.\n\n## Connecting\n1. Open Settings → Google Search Console.\n2. Click **Connect Google Search Console**.\n3. Sign in to Google and grant the read-only scope.\n4. RankMeFast matches the connected property to your site.\n\n## Disconnecting or revoking\nClick **Disconnect** in the same panel and confirm. That deletes our stored token immediately. You can also revoke us at any time from https://myaccount.google.com/permissions.\n\n## \"Reconnect needed\"\nIf Google invalidates our token — for instance, when you change your account password — the card shows a red **Reconnect needed** banner. The audit rules fall back to \"insufficient data\" until you reconnect.\n\n[Back to the docs index](./index.en.md)"},{"slug":"gsc-generative-appearance","locale":"en","title":"Generative AI appearance in Search Console","description":"A first-party GSC signal — how often your site was shown inside Google generative AI features.","section":"research","order":4,"searchText":"# Generative AI appearance in Search Console\n\nGoogle Search Console reports how often your pages appeared inside Google's generative-AI surfaces (for example, AI Overviews). This card shows those numbers verbatim — we do not add, invent, or estimate anything Google did not send.\n\n## What we read\n\nWe call the Search Console API with `dimensions=['searchAppearance']` over the last 28 days ending three days ago (Google's own reporting lag). Every row Google returns is classified from primary Google documentation:\n\n- Known generative-AI values map to a stable label.\n- Unknown values pass through as `Other` with the raw name preserved for support review — never silently reclassified as generative.\n\n## States\n\n- **Available** — Google returned at least one recognized generative-AI row. The card shows the exact clicks, impressions, CTR, and average position Google sent, per appearance row.\n- **Unavailable** — Google returned no recognized generative-AI row for this property and window. This is different from a zero — it means Google did not report the surface at all. We do not zero-fill.\n- **Partial** — Google returned rows but flagged the response as partial (rate limit, truncation). The rows we did receive are still displayed.\n- **Reconnect required** — the Google connection needs to be reconnected. Open the site's Google workspace and follow the reconnect prompt.\n\n## Separate from provider metrics\n\nGenerative-AI appearance is a **first-party** Google signal. It is displayed on the Google workspace next to your other Search Console cards. It is never merged with the mention and share-of-voice charts on the AI Visibility workspace — those come from a different provider and measure a different thing.\n\n## What this does not tell you\n\nSearch Console reports impressions and clicks — how often Google chose to include your page inside a generative surface. It does not tell you whether the AI answer named you, linked you, quoted you, or preferred a competitor. For that, use [AI Visibility](./ai-visibility.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"audience-research","locale":"en","title":"Audience Research","description":"Turn topics, competitors, and a market into audience segments from public evidence.","section":"research","order":5,"searchText":"# Audience Research\n\nAudience Research turns your market, a few seed topics, and optional competitor domains into audience segments — built only from public evidence.\n\n## What you provide\n\n- Your site's market (country and language).\n- Up to ten seed topics you want to explore.\n- Up to five competitor domains (optional).\n\n## Preview before you spend\n\nThe preview shows what a run will use and how many runs you have left this month. It reserves nothing — previewing is always free.\n\n## What a run costs\n\nOne run consumes one `audience_research_runs` unit, no matter how many pages it reads inside. Monthly runs per plan: Free 0, Starter 2, Pro 10, Agency 30. There is no credit pack for this metric — at the cap, wait for the monthly reset or upgrade.\n\nStarting the identical run twice does not charge twice — the second start returns the run already in flight. A later fresh run is a new run and uses a new unit.\n\n## A fixed per-run cost ceiling\n\nEvery run has a fixed per-run cost ceiling, with a stricter inner ceiling for the AI clustering step. A run that reaches the ceiling stops early and keeps the evidence it already collected — the result is labeled **partial results**, never silently truncated.\n\n## How pages are selected\n\nSelection is deterministic — the same inputs pick the same public pages. We rank candidates by topic relevance, then search position, then recency, and spread picks across source types and domains so no single site can dominate the evidence.\n\n## Run states\n\nA run moves through **queued → discovering → selecting → collecting → clustering** and ends as **completed**, **partial results**, or **failed**.\n\n## Refunds\n\nA unit is returned exactly once, and only when a run fails with no usable public evidence at all. A completed or partial run keeps its unit — partial evidence is still evidence.\n\n## Nothing is published\n\nResults are for your account only. A run never publishes, posts, or shares anything anywhere.\n\n## Where results go\n\nSignals you accept appear in [Next Actions](./next-actions.en.md) and can create Content Intelligence briefs. Nothing moves without your decision.\n\n[Back to the docs index](./index.en.md)"},{"slug":"keyword-intelligence","locale":"en","title":"Keyword Intelligence","description":"Gap, overview, trends, and clustering — what each check costs and what the labels mean.","section":"research","order":6,"searchText":"# Keyword Intelligence\n\nKeyword Intelligence is the workspace for keyword gaps, overviews, trends, and AI clustering. It is available on Pro and Agency plans.\n\n## What each check costs\n\n- **Gap** — compares your domain against competitors. One `keyword_lookups` unit per competitor in the comparison (duplicates are removed first).\n- **Overview** — one `keyword_lookups` unit per keyword you look up.\n- **Trends** — one `keyword_lookups` unit per keyword.\n\nA result served from cache still counts as one unit. Caching keeps vendor costs (and your price) down; a check is still a check.\n\n## Reading the labels\n\nEvery number carries a label saying where it came from:\n\n- **Estimate** — derived by simple arithmetic from provider data, such as a trend line computed from monthly volumes.\n- **Provider observation** — a value the data provider reported; we display it as received.\n- **AI interpretation** — produced by an AI step, such as cluster names. Treat it as a suggestion, not a measurement.\n\n## Clustering\n\nClustering groups a keyword list into topics with one AI step. One clustering run uses one `ai_summaries` unit; rerunning the identical list returns the stored result for free.\n\n## When keyword lookups run out\n\nThe `keywords-1k` credit pack adds 80 `keyword_lookups` for $19, one-time. It tops up gap, overview, and trends reads after your monthly allowance is spent — see [Plans, limits & credits](./plans-limits-credits.en.md).\n\n## From clusters to work\n\nAccept a cluster and it routes onward — into a Content Intelligence brief or into your SEO work, following the route suggested on the cluster. Dismissing records the decision and changes nothing else. Every decision is kept.\n\n[Back to the docs index](./index.en.md)"},{"slug":"link-intelligence","locale":"en","title":"Link Intelligence","description":"Run deeper backlink checks and compare link gaps without guessing what a unit covers.","section":"research","order":7,"searchText":"# Link Intelligence\n\nUse Link Intelligence when the standard backlink overview is not detailed enough. It can pull referring domains, anchors, history, or bulk domain ranks, and it can compare your site with up to three competitors.\n\n\n\n## Know what one unit covers\n\n- One submitted deep pull uses one `link_intel_checks` unit.\n- A link-gap run uses one unit for each competitor. Comparing three competitors uses three units.\n- A preview shows the exact unit breakdown but spends nothing. Confirming the run checks your allowance again.\n\nSee [Pricing](./pricing.en.md) for current plan allowances and pack prices. The deep tools require an eligible plan; opening a result you already created is a stored-data read.\n\n## Understand cache and refunds\n\nA result served from cache still counts. A successful check with no rows also counts: it confirms that the source returned no matching links at that time.\n\nIf the source fails before RankMeFast retains any row, that failed unit is returned once. In a multi-competitor gap run, each competitor is settled separately, so useful legs remain charged while a zero-evidence failed leg is refunded. A preview, cancelled preview, or stored-result read never uses a unit.\n\n## Read the result honestly\n\nDeep results carry the time they were observed. A missing first-seen date stays unavailable rather than being invented. Gap percentages are computed from the stored rows, and competitor input order is preserved.\n\nThe checks are **on demand**, not continuous monitoring. Run a new check when you need a fresher observation. Turning the feature off stops new previews and runs, but stored results remain readable.\n\n## Start a check\n\nOpen a site's Backlinks workspace. Choose **Domains**, **Anchors**, **History**, or **Gap**, review the spend preview, then confirm. Use the stored-runs list to reopen a prior result without spending again.\n\nFor monthly caps, packs, and the common refund rule, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"traffic-insights","locale":"en","title":"Traffic Insights","description":"Compare modeled competitor traffic while keeping estimates, units, and refunds clear.","section":"research","order":8,"searchText":"# Traffic Insights\n\nTraffic Insights creates a stored snapshot of a competitor domain: modeled monthly organic visits, domain rank, keyword count, leading countries, and history. Use it for directional comparison, not as a replacement for the competitor's analytics.\n\n\n\n## Know what one unit covers\n\nOne submitted target-domain snapshot uses one `traffic_snapshots` unit. A multi-domain preview shows one unit for each domain, but each confirmed domain becomes its own snapshot. Previewing or cancelling the preview spends nothing.\n\nCurrent monthly allowances and pack prices are on [Pricing](./pricing.en.md). Opening, filtering, or comparing snapshots you already created is a stored-data read and uses no new unit.\n\n## Treat every number as an estimate\n\nEvery numeric Traffic Insights field is labeled **Estimate**. The values are modeled from a search-data index; they are not first-party visits, conversions, or analytics. Missing values remain unavailable. A partial snapshot can still be useful when one part succeeds, and the interface shows that it is partial.\n\nCompare views align up to five stored snapshots. They do not fill gaps or infer missing country or history values.\n\n## Understand cache and refunds\n\nA cached snapshot still counts as one unit. A successful empty result also counts.\n\nThe snapshot collects several related observations. If all of them fail and RankMeFast retains no result, the unit is returned exactly once. If any usable observation is retained, the snapshot is stored as partial or complete and the unit remains consumed. A transport retry cannot return the same unit twice.\n\n## Choose when to refresh\n\nTraffic Insights runs **on demand**. It does not monitor competitors continuously. Start another snapshot when you need a newer estimate. If an operator pauses new snapshots, your stored list, detail, and comparison views remain readable.\n\nOpen a site's **Traffic** tab, enter a public domain, review the spend preview, and confirm. For caps, packs, and refund rules, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"keyword-trends","locale":"en","title":"Keyword Trends","description":"Explore search-interest direction, momentum, and seasonality with honest estimate labels.","section":"research","order":9,"searchText":"# Keyword Trends\n\nKeyword Trends shows how search interest changes over time. It reports a relative index, year-over-year direction, recent momentum, and seasonal months for up to five keywords in one exploration.\n\n\n\n## Know what one unit covers\n\nOne confirmed exploration uses one `trend_explorations` unit, whether you enter one keyword or five. Previewing the request, cancelling the preview, reopening a stored run, or selecting a related-query follow-up before confirming does not spend a unit. A follow-up you confirm is a new exploration.\n\nSee [Pricing](./pricing.en.md) for current allowances and pack prices.\n\n## Read the index as an estimate\n\nEvery series and readout is labeled **Estimate**. The 0–100 search-interest index is relative to the selected market and time window. It is not absolute monthly search volume, traffic, or a forecast.\n\n- Year-over-year needs at least 56 valid observations.\n- Momentum needs 12 observations and can be up, down, or flat.\n- Seasonality needs 24 distinct months. Without enough history, RankMeFast says so instead of inventing a pattern.\n\nA flat, sparse, or empty successful response is a valid observation, not a failure.\n\n## Understand cache and refunds\n\nA result served from cache still uses one unit. A successful exploration with no series also uses one unit.\n\nIf the source fails before any series is retained, the reserved unit is returned exactly once and the run is marked refunded. Once a usable series is retained, the unit remains consumed even if another readout is unavailable.\n\n## Choose when to explore\n\nKeyword Trends is **on demand**, not continuous monitoring. Run it again when you need a newer view. If new explorations are paused, stored runs stay readable.\n\nOpen **Live trends** from Keyword Research, enter up to five phrases, choose the market and language, review the spend preview, and confirm. For monthly caps and the common refund rule, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"review-intelligence","locale":"en","title":"Review Intelligence","description":"Sync public reviews on demand, read trends and cited themes, and understand exactly when a unit is returned.","section":"research","order":10,"searchText":"# Review Intelligence\n\nReview Intelligence brings public reviews from configured Google, Trustpilot, and Tripadvisor listings into one stored inventory. It calculates rating statistics and can group repeated praise and complaints into themes supported by stored review citations.\n\n\n\n## Know what one unit covers\n\nOne confirmed sync uses one `review_syncs` unit for the whole job. You may select one, two, or three configured sources and choose a depth up to 100 reviews per source; the unit cost does not multiply by source count. The theme pass is included.\n\nPreviewing or cancelling spends nothing. See [Pricing](./pricing.en.md) for current monthly allowances and pack prices.\n\n## Understand cache, deduplication, and refunds\n\nPublic review material served from cache still counts. RankMeFast deduplicates by source and source review ID, so a successful sync that finds no new rows also counts: the source was checked and existing reviews were confirmed.\n\nThe unit is returned exactly once only when **every selected source** ends in a source failure and no new review row is retained. A partial sync, an empty or fully deduplicated success, or a theme-generation failure still consumes the unit because review work succeeded or evidence remains. Retrying cannot return the same unit twice.\n\n## Read themes and statistics\n\nRatings, counts, source splits, and monthly trends come from stored review rows. Generated themes are optional. A theme is shown only when it cites stored reviews; unreliable theme text is dropped. If theme generation fails, your reviews and deterministic statistics remain available.\n\nSearch, filter, open a stored run, and export matching rows as CSV without using another sync unit. Review text is displayed as text, not executable markup.\n\n## Choose when to sync\n\nReview Intelligence is **on demand**. It does not monitor listings continuously, send review alerts, reply to reviews, or draft responses. Start a new sync when you want a fresher inventory. Pausing new syncs does not hide stored sources, runs, reviews, statistics, themes, or exports.\n\nOpen a site's **Reviews** tab, configure at least one source, review the spend preview, and confirm. For caps and refunds, see [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"brand-radar","locale":"en","title":"Brand Radar","description":"Scan one brand query on demand, inspect retained mentions, and understand the add-on and refund boundary.","section":"research","order":11,"searchText":"# Brand Radar\n\nBrand Radar searches for public mentions of one brand query, stores the retained rows, and builds counts, sentiment distribution, top domains, and a comparison with the previous settled scan for the same query. It can also create a short cited digest.\n\n\n\n## Know what one unit covers\n\nOne confirmed scan of one brand query uses one `brand_mention_scans` unit. The optional language/location filters and the cited digest are included. A scan cannot bundle competitor queries into the same unit.\n\nPreviewing or cancelling spends nothing. The preview always says a fresh fetch is required because mention rows are account-scoped and are not shared through a cross-account cache. Current base allowances, the optional Brand Radar add-on, and pack prices are on [Pricing](./pricing.en.md).\n\n## Understand refunds and partial results\n\nThe unit is returned exactly once only if the initial mention search ends in a source failure and RankMeFast retains no mention row.\n\nThese outcomes still consume the unit:\n\n- the search succeeds but finds no mentions;\n- some mentions are retained before a later source or budget stop;\n- the summary or generated digest fails;\n- every generated digest sentence is dropped because its citation cannot be verified.\n\nRetained evidence is never discarded just to create a refund.\n\n## Separate measured facts from generated prose\n\nMention count, sentiment split, top domains, and change from the previous same-query scan are computed from stored rows. No previous scan means “no comparison yet,” never a made-up zero.\n\nDigest sentences are generated prose and are shown only with citations to retained mention rows. If no reliable sentence survives, the interface says so. You can inspect mention rows and export them as CSV.\n\n## Choose when to scan\n\nBrand Radar is **on demand**, not continuous monitoring. Start a new scan when you need a new observation. Weekly Pulse may summarize deltas already stored by Brand Radar, but it does not start a scan or spend another scan unit. If new scans are paused, stored lists, details, citations, trends, and exports remain readable.\n\nFor eligibility, allowances, and refund rules, read [Plans, limits & credits](./plans-limits-credits.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"serp-sensor","locale":"en","title":"SERP Sensor methodology","description":"How the public daily volatility index is collected, calculated, budgeted, and limited.","section":"research","order":12,"searchText":"# SERP Sensor methodology\n\nThe [SERP Sensor](/serp-sensor) is a shared public index of observed search-result movement. It is not a personal rank tracker, a forecast, or a monitor for your own keyword set.\n\n\n\n## Fixed basket and cadence\n\nRankMeFast checks the same 200 fixed US-English keywords once per UTC day: 20 keywords in each of 10 categories. For every keyword it keeps the first 20 unique public HTTP or HTTPS result URLs. Users cannot change the keywords, market, language, depth, or collection time.\n\nThe categories cover finance, health, travel, technology, retail, entertainment, local services, education, real estate, and automotive. The fixed basket makes one day comparable with the next; it does not represent every search or location.\n\n## How the index is calculated\n\nFor one keyword, each result's current rank is compared with its previous rank. A result missing on one day is treated as rank 21. The absolute rank movements across the union of both result sets are added and divided by 420, the maximum displacement for a top-20 set. This produces a value from 0 to 1.\n\nRankMeFast averages the 20 keyword values in a category and scales the result to 0–10, rounded to one decimal place. The composite is the average of the category values that were published that day. The labels describe observed displacement:\n\n| Index | Label |\n|---|---|\n| 0.0–2.0 | Low |\n| above 2.0–4.5 | Routine |\n| above 4.5–7.0 | Elevated |\n| above 7.0–10.0 | High |\n\nThe first complete observation is a **baseline** and has no numeric change. If only some categories complete, the composite is marked partial. Missing data is not estimated or filled.\n\n## Budget and availability\n\nThe sensor uses a global operator budget, not a customer allowance. It consumes no account unit, has no customer cache rule, and has no refund because customers are never charged. The collector checks its budget before and during the daily run. A category skipped for budget or collection failure is labeled accordingly.\n\nPublic history contains at most the latest 30 stored dates per category. Turning the sensor off returns an unavailable state without exposing the fixed keywords or result URLs.\n\nFor customer-run research features and their units, see [Pricing](./pricing.en.md).\n\n[Back to the docs index](./index.en.md)"},{"slug":"pricing","locale":"en","title":"Pricing","description":"How RankMeFast plans and monthly allowances work.","section":"account","order":1,"searchText":"# Pricing\n\nRankMeFast sells you a monthly allowance of SEO checks. Free forever tier for kicking the tyres; Starter and Pro for founders and teams; Agency for client work.\n\n## What we count\n\nEvery plan meters the same units:\n\n- **Audit** — one full audit run over a site.\n- **Audit page** — one page inside that run (crawl budget).\n- **Tracked keyword** — one keyword you keep watching.\n- **Rank check** — one SERP fetch for a tracked keyword; the weekly sweep and every manual re-check count.\n- **Keyword lookup** — one site-keyword discovery request or one read from the full keyword-research panel.\n- **Backlink row** — one row on the backlink list.\n- **AI summary** — one Claude-written plain-language recap of the fix-now list.\n- **Competitor lookup** — one competitor list or intersection call.\n\nCached results still count as one check. Caching is how we keep vendor costs (and your price) down; a check is still a check.\n\nMonthly keyword lookups are **5 Free / 20 Starter / 50 Pro / 100 Agency**. Site discovery is included on every plan; full keyword research remains Pro and Agency.\n\n## New intelligence units\n\n\n- **Keyword Trends exploration** — one confirmed exploration of up to five phrases.\n- **Traffic Insights snapshot** — one confirmed target domain.\n- **Link Intelligence check** — one deep pull, or one unit for each competitor in a link-gap run.\n- **Review Intelligence sync** — one whole sync of one to three configured sources.\n- **Brand Radar scan** — one brand query; its cited digest is included.\n\nThese five features run on demand, not as continuous monitoring. Cache hits count, and successful empty results count. A reserved unit is returned exactly once only when the feature's source fails before it retains any usable evidence; each guide explains the narrower multi-part rule. Brand Radar always makes a fresh account-scoped fetch rather than sharing mention rows between accounts.\n\nRead the guides for [Link Intelligence](./link-intelligence.en.md), [Traffic Insights](./traffic-insights.en.md), [Keyword Trends](./keyword-trends.en.md), [Review Intelligence](./review-intelligence.en.md), and [Brand Radar](./brand-radar.en.md).\n\n## Brand Radar add-on and intelligence packs\n\n\n\nThe optional Brand Radar add-on is available on Pro and Agency. It adds 60 scans each month for $19/month. Agency already includes 20 monthly scans; Pro needs the add-on to start scanning.\n\nOne-time intelligence packs stay on your balance until used:\n\n- 40 Brand Radar scans for $29.\n- 50 Link Intelligence checks for $19.\n- 50 Review Intelligence syncs for $19.\n- 100 Traffic Insights snapshots for $19.\n- 200 Keyword Trends explorations for $19.\n\n## What happens at a cap\n\n1. The app warns you before you hit a cap.\n2. New requests hard-stop at the cap. Nothing is billed beyond the ceiling.\n3. Three ways out: wait for the monthly reset, buy a credit pack (+80 keyword lookups for $19, or +5 audits for $29 — one-time, never expire), or upgrade your plan.\n\nThe keyword pack is the `keywords-1k` pack: +80 `keyword_lookups` for $19, one-time. It tops up keyword gap, overview, and trends reads when your monthly allowance runs out. Audience Research runs have no credit pack — upgrade or wait for the reset.\n\n## Yearly billing\n\nYearly gives you two months free — 20% off the monthly price. Switch between monthly and yearly anytime from the billing page.\n\n## Cancelling\n\nCancel from the billing page. No calls, no retention gauntlet. Your audits and rank history remain read-only for 30 days."},{"slug":"plans-limits-credits","locale":"en","title":"Plans, limits & credits","description":"Starter, Pro, Agency — caps, trial, overage credits.","section":"account","order":2,"searchText":"# Plans, limits & credits\n\nRankMeFast has three subscription tiers and a free one-off audit.\n\n## Free\n- One full audit at sign-up. No credit card. No ongoing checks.\n\n## The three tiers\n\n\n| Plan | Monthly (USD) | Yearly (USD) | Sites | Keywords | Audits/month | Audit pages | Backlink rows | AI summaries | Seats | Audience Research runs |\n|----------|---------------|--------------|-------|----------|--------------|-------------|---------------|--------------|-------|------------------------|\n| Starter | $49 | $470 | 2 | 250 | 10 | 1,000 | 0 | 20 | 1 | 2 |\n| Pro | $149 | $1,430 | 5 | 1,000 | 30 | 3,000 | 10,000 | 100 | 3 | 10 |\n| Agency | $479 | $4,598 | Unlimited | 2,000 | 45 | 5,000 | 100,000 | 500 | Unlimited | 30 |\n\nPrices above are in USD. Annual is 20 % off the monthly total. Daily rank tracking is a paid add-on on every plan — never bundled.\n\n## Audience Research runs\n\n\nOne run is one whole research job, no matter how many pages it reads. Free accounts get 0 runs; Starter 2, Pro 10, Agency 30 per month. There is no credit pack for this metric — at the cap, wait for the monthly reset or upgrade. See [Audience Research](./audience-research.en.md).\n\n## Keyword Intelligence and Weekly Pulse units\n\n\n- A keyword gap check uses one `keyword_lookups` unit per competitor you compare.\n- A keyword overview or trends read uses one `keyword_lookups` unit per keyword. Results served from cache still count.\n- Clustering a keyword list uses one `ai_summaries` unit; rerunning the identical list is free.\n- One Weekly Pulse digest uses one `ai_mentions_checks` unit per site per week, no matter how many teammates receive it.\n\nDetails live in [Keyword Intelligence](./keyword-intelligence.en.md) and [Weekly Pulse](./weekly-pulse.en.md).\n\n## Intelligence allowances and units\n\n\n| Plan | Keyword Trends explorations | Traffic snapshots | Link Intelligence checks | Review syncs | Brand Radar scans |\n|---|---:|---:|---:|---:|---:|\n| Free | 0 | 0 | 0 | 0 | 0 |\n| Starter | 10 | 5 | 0 | 0 | 0 |\n| Pro | 40 | 25 | 25 | 10 | 0 |\n| Agency | 80 | 80 | 80 | 50 | 20 |\n\nOne Trends unit covers up to five phrases. One Traffic unit covers one target domain. One Link unit covers one deep pull or one competitor leg in a gap run. One Review unit covers a whole one-to-three-source sync. One Brand Radar unit covers one brand query and its cited digest.\n\nCache hits and successful empty results count. A unit is returned exactly once only when the source fails before the feature retains usable evidence. Multi-part Link, Traffic, and Review jobs apply that rule to their documented legs; see [Link Intelligence](./link-intelligence.en.md), [Traffic Insights](./traffic-insights.en.md), [Keyword Trends](./keyword-trends.en.md), [Review Intelligence](./review-intelligence.en.md), and [Brand Radar](./brand-radar.en.md).\n\n## Brand Radar add-on and intelligence packs\n\n\n\nPro and Agency can add 60 Brand Radar scans per month for $19/month. Agency's 20 base scans remain available without the add-on; Pro's base is zero.\n\nThe one-time packs are 40 Brand Radar scans for $29, 50 Link checks for $19, 50 Review syncs for $19, 100 Traffic snapshots for $19, and 200 Trends explorations for $19. Purchased pack balances remain until used. Current purchase options are always shown on [Pricing](./pricing.en.md) and under **Billing → Credits**.\n\n## 14-day free trial\nEvery paid tier includes a 14-day trial. You can cancel any time during the trial without a charge. Card details are collected at sign-up; the first payment lands on day 15.\n\n## What happens at a cap\nWhen you hit a monthly cap (audits, keywords tracked, backlink rows, AI summaries) the specific action is blocked with a message that points at:\n1. Upgrading to a higher tier.\n2. Buying an **overage credit pack** for the matching metric.\n\nNothing else on the account changes — your existing sites, keywords, and reports stay intact.\n\n## Overage credit packs\nSmall one-time packs of extra usage priced per metric. They top up the matching cap and remain on the account until used. Existing subscribers see them under Billing → Credits.\n\n## Cancelling and downgrading\nCancelling keeps your access until the end of the billing period. Downgrading takes effect at the next renewal — your extra sites and keywords are not deleted, they just become read-only until you delete some or upgrade again.\n\n[Back to the docs index](./index.en.md)"},{"slug":"settings-security","locale":"en","title":"Account security","description":"Change your password and manage account credentials.","section":"account","order":3,"searchText":"# Account security\n\nManage the credentials that protect your RankMeFast account from **Settings → Security** (URL: `/settings/security`).\n\n## Change your password\n\n1. Sign in and open **Settings → Security**.\n2. Fill in your **current password**, then choose a **new password** and confirm it. The new password must be at least eight characters and must differ from the current one.\n3. Click **Update password**.\n\nWe use scrypt to store password hashes — we never see your password in plain text. When you change your password successfully:\n\n- Every **other** active session for your account is signed out. The device you used to change the password stays signed in.\n- We send a confirmation email to the address on your account. If you did not change the password, that email is your signal to reset your password immediately via [the forgot-password flow](./getting-started.en.md).\n\nIf the current password you entered is wrong, we say so inline — we never reveal whether an account exists for a given email.\n\n## Change your email\n\n1. Sign in and open **Settings → Security**.\n2. Enter the **new email** address in the *Change email* card. It must be different from your current one.\n3. Click **Send verification link**.\n\nWe send a confirmation link to the **new** address. Your account email only changes after you follow that link:\n\n- Your **current email keeps working** for sign-in until you confirm.\n- If the new address is already registered to another account, we reject the change with a generic *email unavailable* message — we never reveal who owns that address.\n- The link is single-use and expires after a short window; request another change from the same page if it lapses.\n\n## Forgot your current password?\n\nIf you cannot remember your current password, sign out and use **Forgot password?** on the login screen. The emailed link takes you to a page where you can pick a new password without confirming the old one.\n\n\n## Two-factor authentication\n\nAdd a second step at sign-in so a stolen password alone cannot get into your account.\n\n### Enable two-factor\n\n1. Sign in and open **Settings → Security**.\n2. In the **Two-factor authentication** card, enter your current password and click **Enable two-factor**.\n3. Scan the QR code with an authenticator app like 1Password, Authy, or Google Authenticator — or paste the manual-entry code into the app.\n4. Copy or download the **backup codes**. Each code works exactly once. They are the only way back into your account if you lose your device.\n5. Enter the six-digit code your authenticator app is showing right now and click **Verify and turn on**.\n\nOnce enabled, every sign-in asks for a fresh code from your authenticator (or a one-time backup code).\n\n### Disable two-factor\n\nOpen the same card, click **Disable two-factor**, and enter your current password. We turn the feature off immediately.\n\n### Admin accounts\n\nIf your account has the **Admin** role, two-factor authentication is required. You will not be able to open the admin panel until you enable it from **Settings → Security**.\n\n### Lost your device\n\nUse one of your backup codes on the sign-in challenge screen — the same screen has a **Use a backup code** link. Each code works once; if you run out, an admin on your team can reset the second factor from the admin panel.\n\n## Your data rights\n\nYou can request an export or a deletion of your account data. Both cover everything the research surfaces store for your account: rank observations, AI Visibility citations and source gaps, Audience Research runs and their evidence, and cached keyword results. Deletion respects a grace period before the purge — you can cancel the request during that window."},{"slug":"notification-preferences","locale":"en","title":"Notification preferences","description":"Choose which non-security emails you want to receive.","section":"account","order":4,"searchText":"# Notification preferences\n\nYou control which non-security emails RankMeFast sends you. Open **Settings → Notifications** (`/settings/notifications`) and toggle any of the four switches.\n\nEvery switch defaults to **on**. Turning a switch off takes effect immediately — the next matching email is skipped.\n\n## What each toggle controls\n\n- **Audit complete** — Sent when an audit run finishes for one of your sites.\n- **Rank drops** — Sent when a tracked keyword loses positions in your latest rank check.\n- **Billing** — Payment receipts, failed-payment reminders, and dunning notices. Turning this off silences ALL billing email; you can still see receipts and past invoices in **Billing → Invoices**.\n- **Product updates** — Occasional welcome messages, product news, and tips.\n\n## Weekly Pulse is a separate opt-in\n\nThe [Weekly Pulse](./weekly-pulse.en.md) digest is not controlled by the switches above. Each teammate turns it on per site from the site's AI Visibility workspace, and it stays off until they do.\n\nDelivery retries are safe: a retry can never send the same digest twice. If your address cannot receive the email, other subscribed teammates still get theirs.\n\n## Security email cannot be disabled\n\nThe following emails are always sent, regardless of the switches above:\n\n- Password reset link.\n- Email address verification.\n- Password-changed confirmation.\n\nThose emails exist to protect your account. If you receive one you did not expect, treat it as your signal to review your active sessions and reset your password from **Settings → Security**."},{"slug":"team","locale":"en","title":"Team & invites","description":"Invite teammates, accept invites, and manage seats on your plan.","section":"account","order":5,"searchText":"# Team & invites\n\nEvery RankMeFast account can share access with additional teammates. You manage members and pending invites from **Settings → Team**.\n\n## Seats per plan\n\n| Plan | Included seats |\n|------|----------------|\n| Free | 1 |\n| Starter | 1 |\n| Pro | 3 |\n| Agency | Unlimited |\n\nAn account's owner always counts as one seat. The current usage is shown at the top of the page (for example, \"2 of 3 seats used\"). When every seat is taken, the invite form is disabled and an **Upgrade plan** link appears.\n\n## Invite a teammate\n\n1. Open **Settings → Team**.\n2. Under **Invite a teammate**, enter your colleague's email address.\n3. Click **Send invite**.\n\nWe email a single-use link that is valid for seven days. The pending invite immediately reserves a seat.\n\n## Accept an invite\n\nClick the link in the email — or open `/team/accept/<token>` in your browser while signed in with the invited email. Once accepted, the seat converts from **Pending** to **Active** and the teammate gains access to the workspace.\n\n- If the link is opened by a different account than the one it was addressed to, RankMeFast responds with a \"not found\" error — invite tokens are single-use and can only be redeemed by the invited email.\n- If the link has expired, ask the workspace owner to send a fresh invite.\n- If the invite has already been accepted, opening the link again returns \"already accepted\".\n\n## Remove a teammate or revoke a pending invite\n\n1. Open **Settings → Team**.\n2. Open the ⋯ menu on the member's row.\n3. Choose **Remove** for an accepted member, or **Revoke invite** for a pending one.\n\nRemoving a member (or revoking a pending invite) frees the seat immediately — you can invite someone else without waiting.\n\nThe workspace owner cannot be removed from the team. Team members can remove themselves (a \"leave\" action)."},{"slug":"public-api","locale":"en","title":"Public API","description":"Read your sites, reports, rankings, and keywords over a simple key-authenticated API. Agency feature.","section":"developers","order":1,"searchText":"# Public API\n\nThe public API gives you read-only access to the data RankMeFast already holds for your account — sites, the latest audit report, rank history, and tracked keywords. It is an **Agency** feature (see [Plans, limits & credits](./plans-limits-credits.en.md)). The API never triggers new vendor work — it only reads what your audits and rank checks have already produced.\n\n## Authentication\n\nCreate a key under **Account → API keys** (`/profile?tab=api-keys`). The full key is shown **exactly once** — copy it right away; afterwards only its prefix is visible. You can hold up to ten active keys and revoke any of them at any time. A revoked key stops working immediately.\n\nSend the key as a bearer token on every request:\n\n```\nAuthorization: Bearer rmf_XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX\n```\n\nReplace `https://your-rankme-host` in the examples below with your api origin (the `SERVER_URL` of your install).\n\n## Endpoints\n\n### List your sites\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n https://your-rankme-host/api/v1/sites\n```\n\nReturns `{ \"sites\": [{ \"id\", \"domain\", \"url\", \"createdAt\" }] }`.\n\n### Latest audit report for a site\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n https://your-rankme-host/api/v1/sites/<siteId>/report/latest\n```\n\nReturns the most recent **succeeded** audit as `{ \"runId\", \"report\" }` — the same findings, buckets, and localized copy the dashboard shows. Responds `404` when the site has no finished audit yet.\n\n### Rank history for a site\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n \"https://your-rankme-host/api/v1/sites/<siteId>/rank-history?from=2026-06-01&to=2026-07-01\"\n```\n\nReturns `{ \"keywords\": [{ \"id\", \"phrase\", \"series\": [...] }] }`. Each series point carries the position, the URL that ranked, and the Google AI Overview signals (`aiOverviewPresent`, `aiCited`, `aiCitedUrl`). `from` and `to` are optional ISO dates.\n\n### All tracked keywords\n\n```bash\ncurl -H \"Authorization: Bearer rmf_...\" \\\n https://your-rankme-host/api/v1/keywords\n```\n\nReturns every tracked keyword across your sites with its latest position, delta, and AI Overview fields.\n\n## Compatibility\n\nThis release exposes exactly four read-only routes:\n\n- `GET /api/v1/sites`\n- `GET /api/v1/sites/:siteId/report/latest`\n- `GET /api/v1/sites/:siteId/rank-history`\n- `GET /api/v1/keywords`\n\nBrand Radar, Review Intelligence, Link Intelligence, Traffic Insights, Keyword Trends, and the SERP Sensor do not add routes under `/api/v1`. The five paid surfaces remain authenticated dashboard features; the public sensor remains under `/api/volatility`. Existing response fields keep their meaning, and clients should ignore additive fields they do not recognize.\n\n\n\n## Rate limits\n\nBy default each key may make **120 requests per minute**. Above that the API answers `429` until the window resets.\n\n## Errors\n\nErrors use the shape `{ \"error\": { \"message\": \"...\", \"details\": ... } }`, localized by your `Accept-Language` header:\n\n- `401` — the key is missing, malformed, revoked, or unknown.\n- `402` — your plan does not include the API.\n- `404` — the site or report does not exist on your account.\n- `429` — rate limit exceeded (body: `{ \"error\": \"...\" }`).\n\n[Back to the docs index](./index.en.md)"},{"slug":"rankmefast-mcp","locale":"en","title":"RankMeFast MCP","description":"Connect RankMeFast to coding agents, IDEs, and MCP-ready CLIs.","section":"developers","order":2,"searchText":"# Connect RankMeFast to your AI tools\n\nRankMeFast has a remote [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) endpoint. It lets a compatible coding agent, IDE, or CLI read your SEO data and start audits without leaving your current workflow.\n\n## Set up in five minutes\n\n1. Open [Settings → API keys](/profile?tab=api-keys), select **Create key**, and copy the key when it appears. RankMeFast shows the full key only once.\n2. Use the RankMeFast MCP endpoint:\n\n ```text\n https://rankme.fast/api/mcp\n ```\n\n3. For clients that support environment variables, set the key before starting the client:\n\n ```bash\n export RANKMEFAST_API_KEY='rmf_REPLACE_WITH_YOUR_KEY'\n ```\n\n In PowerShell:\n\n ```powershell\n $env:RANKMEFAST_API_KEY = 'rmf_REPLACE_WITH_YOUR_KEY'\n ```\n\n4. Choose your client below, paste its configuration, and restart or reload the client.\n5. Verify it with: **“Use RankMeFast to list my sites.”**\n\n> Treat the key like a password. RankMeFast API keys are account-wide and do not have separate scopes. The `start_audit` tool can consume your monthly audit allowance. Prefer an environment variable, password prompt, or user-level config, and never commit a key to your repository.\n\nMCP access is included with Starter, Pro, and Agency. Agency keys also work with the read-only public REST API.\n\n## Choose your client\n\n| Client or surface | Direct setup | Configuration |\n| --- | --- | --- |\n| Claude Code and its desktop Code tab | Yes | HTTP server with bearer header |\n| Cursor IDE and Cursor Agent CLI | Yes | User or project MCP JSON |\n| VS Code with GitHub Copilot | Yes | User or workspace `mcp.json` |\n| GitHub Copilot CLI | Yes | CLI command or user JSON |\n| Windsurf / Cascade | Yes | User MCP JSON |\n| Codex CLI, IDE extension, and ChatGPT desktop Codex | Yes | Shared Codex TOML |\n| Gemini CLI | Yes | CLI command or user JSON |\n| OpenCode | Yes | Remote MCP JSON |\n| JetBrains AI Assistant and Junie | Yes | IDE MCP settings |\n| Zed | Yes | Remote context server |\n| Cline | Yes | Streamable HTTP server |\n| Roo Code | Yes | Streamable HTTP server |\n| Kiro IDE and CLI | Yes | User or project MCP JSON |\n| Copilot in Visual Studio, JetBrains, Xcode, and Eclipse | Yes | Copilot MCP JSON |\n| Claude.ai / Claude Desktop chat connector | Not directly | The connector requires OAuth; RankMeFast currently uses bearer keys |\n| ChatGPT web | Not directly | It does not read local Codex MCP configuration |\n\nAny other client can connect if it supports remote **Streamable HTTP** and a custom `Authorization` header. RankMeFast does not expose a local stdio or legacy SSE server.\n\n## Claude Code\n\nAdd a user-level server. The single quotes keep the environment variable reference intact:\n\n```bash\nclaude mcp add-json --scope user rankmefast \\\n '{\"type\":\"http\",\"url\":\"https://rankme.fast/api/mcp\",\"headers\":{\"Authorization\":\"Bearer ${RANKMEFAST_API_KEY}\"}}'\n```\n\nCheck the connection with:\n\n```bash\nclaude mcp get rankmefast\n```\n\nYou can also run `/mcp` inside Claude Code. The Claude Code surface in the desktop app uses the same configuration. See the official [Claude Code MCP guide](https://code.claude.com/docs/en/mcp).\n\nThe Claude.ai and Claude Desktop **chat connector** is different: its remote connector flow documents OAuth, not an arbitrary static bearer header. It cannot connect directly until RankMeFast offers OAuth; use Claude Code instead. See [Claude custom connectors](https://support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp).\n\n## Cursor IDE and Cursor Agent CLI\n\nCreate `~/.cursor/mcp.json` for a user-level setup. Replace both placeholders; keep this file private:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n}\n```\n\nCursor IDE and Cursor Agent CLI read the same configuration. Verify it with:\n\n```bash\ncursor-agent mcp list\ncursor-agent mcp list-tools rankmefast\n```\n\nFor a team configuration, use `.cursor/mcp.json` but do not put a literal key in that committed file. See the official [Cursor MCP guide](https://docs.cursor.com/context/model-context-protocol).\n\n## VS Code and GitHub Copilot\n\nRun **MCP: Open User Configuration** from the Command Palette, then use a password input so the secret is not written into the file:\n\n```json\n{\n \"inputs\": [\n {\n \"type\": \"promptString\",\n \"id\": \"rankmefast-key\",\n \"description\": \"RankMeFast API key\",\n \"password\": true\n }\n ],\n \"servers\": {\n \"rankmefast\": {\n \"type\": \"http\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer ${input:rankmefast-key}\"\n }\n }\n }\n}\n```\n\nRun **MCP: List Servers** to start or inspect it. Workspace configuration can live in `.vscode/mcp.json`; the password-input form is safe to share. See the official [VS Code MCP configuration reference](https://code.visualstudio.com/docs/agents/reference/mcp-configuration).\n\n## GitHub Copilot CLI\n\nWith `RANKMEFAST_API_KEY` set in your shell:\n\n```bash\ncopilot mcp add \\\n rankmefast \\\n --type http \\\n --url https://rankme.fast/api/mcp \\\n --header \"Authorization=Bearer $RANKMEFAST_API_KEY\" \\\n --tools \"*\"\n```\n\nThe shell expands the key before Copilot saves the user configuration, so protect `~/.copilot/mcp-config.json`. See [Adding MCP servers to Copilot CLI](https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/add-mcp-servers).\n\n## Windsurf / Cascade\n\nAdd this to `~/.codeium/windsurf/mcp_config.json`:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"serverUrl\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer ${env:RANKMEFAST_API_KEY}\"\n }\n }\n }\n}\n```\n\nReload it from **Windsurf Settings → Cascade → MCP Servers**. See the official [Windsurf MCP guide](https://docs.windsurf.com/windsurf/cascade/mcp).\n\n## Codex CLI, IDE extension, and ChatGPT desktop\n\nCodex CLI, the Codex IDE extension, and the Codex surface in ChatGPT desktop share `~/.codex/config.toml` on the same machine:\n\n```toml\n[mcp_servers.rankmefast]\nurl = \"https://rankme.fast/api/mcp\"\nbearer_token_env_var = \"RANKMEFAST_API_KEY\"\n```\n\nRestart the IDE or desktop app after setting the environment variable. In the CLI, check it with:\n\n```bash\ncodex mcp list\n```\n\nYou can also run `/mcp` in a Codex session. See the official [Codex MCP guide](https://learn.chatgpt.com/docs/extend/mcp).\n\nChatGPT web does not read local Codex configuration, so this setup applies only to Codex surfaces on the configured machine.\n\n## Gemini CLI\n\nWith `RANKMEFAST_API_KEY` set in your shell:\n\n```bash\ngemini mcp add \\\n --scope user \\\n --transport http \\\n --header \"Authorization: Bearer $RANKMEFAST_API_KEY\" \\\n rankmefast https://rankme.fast/api/mcp\n```\n\nVerify it with `gemini mcp list` or `/mcp` inside Gemini CLI. The command stores the expanded header in `~/.gemini/settings.json`, so keep that file private. In JSON, use `httpUrl`; Gemini reserves `url` for legacy SSE. See the official [Gemini CLI MCP guide](https://github.com/google-gemini/gemini-cli/blob/main/docs/tools/mcp-server.md).\n\n## OpenCode\n\nAdd this to the global `~/.config/opencode/opencode.json` file:\n\n```json\n{\n \"$schema\": \"https://opencode.ai/config.json\",\n \"mcp\": {\n \"rankmefast\": {\n \"type\": \"remote\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"enabled\": true,\n \"oauth\": false,\n \"headers\": {\n \"Authorization\": \"Bearer {env:RANKMEFAST_API_KEY}\"\n }\n }\n }\n}\n```\n\n`oauth: false` prevents OAuth discovery because RankMeFast uses a bearer key. Check the setup with:\n\n```bash\nopencode mcp list\nopencode mcp debug rankmefast\n```\n\nSee the official [OpenCode MCP guide](https://opencode.ai/docs/mcp-servers/).\n\n## JetBrains AI Assistant and Junie\n\nOpen **Settings → Tools → AI Assistant → Model Context Protocol (MCP)**, add an HTTP server, and paste:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n}\n```\n\nKeep this in IDE-level settings rather than a project file. Enable **Pass custom MCP servers** when you want Junie or another integrated agent to receive the tools. See the official [JetBrains MCP guide](https://www.jetbrains.com/help/ai-assistant/mcp.html).\n\n## Zed\n\nOpen **Settings → AI → MCP Servers → Add Remote Server**, or add this to your user settings:\n\n```json\n{\n \"context_servers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n}\n```\n\nKeep the literal key in user settings, not project settings. A green indicator beside the server confirms the connection. See the official [Zed MCP guide](https://zed.dev/docs/ai/mcp).\n\n## Cline\n\nOpen Cline's MCP settings and add a Streamable HTTP server:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"type\": \"streamableHttp\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n },\n \"disabled\": false,\n \"autoApprove\": []\n }\n }\n}\n```\n\nLeave `autoApprove` empty so starting an audit requires your approval. See the official [Cline MCP guide](https://docs.cline.bot/mcp/mcp-overview).\n\n## Roo Code\n\nUse the global MCP settings or `.roo/mcp.json`:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"type\": \"streamable-http\",\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n },\n \"disabled\": false,\n \"alwaysAllow\": []\n }\n }\n}\n```\n\nLeave `alwaysAllow` empty so spending actions need approval. See the official [Roo Code MCP guide](https://docs.roocode.com/features/mcp/using-mcp-in-roo).\n\n## Kiro IDE and CLI\n\nUse `~/.kiro/settings/mcp.json` globally or `.kiro/settings/mcp.json` for a project:\n\n```json\n{\n \"mcpServers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer ${RANKMEFAST_API_KEY}\"\n }\n }\n }\n}\n```\n\nKiro expands environment variables in headers. See the official [Kiro MCP configuration](https://kiro.dev/docs/mcp/configuration/).\n\n## Copilot in other IDEs\n\nGitHub Copilot in Visual Studio, JetBrains IDEs, Xcode, and Eclipse uses a different shape from VS Code:\n\n```json\n{\n \"servers\": {\n \"rankmefast\": {\n \"url\": \"https://rankme.fast/api/mcp\",\n \"requestInit\": {\n \"headers\": {\n \"Authorization\": \"Bearer rmf_REPLACE_WITH_YOUR_KEY\"\n }\n }\n }\n }\n}\n```\n\nStore this as user-level configuration and do not commit the key. Follow GitHub's [MCP setup guide for Copilot IDE extensions](https://docs.github.com/en/copilot/how-tos/provide-context/use-mcp-in-your-ide/extend-copilot-chat-with-mcp) for the config-file location in your IDE.\n\n## Available RankMeFast tools\n\n| Tool | What it does | Uses plan allowance? |\n| --- | --- | --- |\n| `list_sites` | Lists sites owned by your account | No |\n| `get_latest_audit_report` | Returns the latest completed audit for a site | No |\n| `list_keywords` | Lists tracked keywords for a site | No |\n| `get_rank_history` | Returns keyword rank history for a site | No |\n| `list_content_analyses` | Lists Content Intelligence analyses | No |\n| `get_content_analysis` | Returns one Content Intelligence analysis | No |\n| `start_audit` | Starts a site audit, within your plan's page cap | **Yes — one audit** |\n| `get_audit_status` | Checks an audit run | No |\n\nReads and actions are always scoped to the account that owns the key. A site, analysis, or run ID from another account is returned as not found.\n\nThe tool list changes additively: existing tools keep their names and argument shapes, and new tools may appear as features ship. Nothing in this release removes or renames a tool.\n\n## Troubleshooting\n\n- **The client cannot discover tools:** confirm the URL ends in `/api/mcp`, choose Streamable HTTP rather than SSE or stdio, then restart the client.\n- **401 Unauthorized:** use the exact `Authorization: Bearer rmf_…` header. The key may be mistyped, revoked, or attached to an inactive account.\n- **402 Upgrade required:** MCP needs Starter, Pro, or Agency. See [Plans, limits & credits](./plans-limits-credits.en.md).\n- **405 Method not allowed in a browser:** this is expected because the endpoint accepts MCP `POST` requests, not ordinary browser `GET` requests.\n- **429 Too many requests:** wait for the rate-limit window shown in the response headers, then retry.\n- **503 Unavailable:** MCP is disabled or temporarily unavailable on that RankMeFast installation; ask its operator.\n- **Works locally but not through a proxy:** make sure the proxy forwards `Authorization`, `Content-Type`, and `Accept` headers to `/api/mcp`.\n\nTo replace a key, create a new one, update and verify every client, then revoke the old key from [Settings → API keys](/profile?tab=api-keys). Revocation is immediate.\n\n## Compatibility\n\nThe registry in this release is exactly the eight tools listed above. Seven are read-only; `start_audit` is the only tool that spends a plan allowance. Brand Radar, Review Intelligence, Link Intelligence, Traffic Insights, Keyword Trends, and the SERP Sensor do not have MCP tools. No existing tool name or argument shape changed.\n\n\n\n## Related guides\n\n- [Public API](./public-api.en.md) — the Agency-only read API.\n- [Content Intelligence](./content-intelligence.en.md) — understand the analysis data returned by MCP.\n- [Account security](./settings-security.en.md) — keep your account and keys safe."}]}},"actionData":null,"errors":null}</article></body></html>