# SerpBeats > SerpBeats is a search visibility platform for tracking daily Google rankings (by country, city and device), competitors, technical SEO health, Google Maps (local) rankings, Google AI Overviews, and brand citations in AI answer engines such as ChatGPT, Gemini and Perplexity. It is built for site owners, SEO professionals, in-house teams and agencies who want search and AI visibility in one dashboard. Customers can also connect their own AI agent (Claude, ChatGPT, Gemini, Codex, Claude Code, Cursor) to read their SerpBeats data through an MCP server. SerpBeats is available in Turkish and English. Public prices are monthly and exclude VAT; tax is added at checkout according to the buyer's country. Visitors identified as being in Turkey see TRY pricing, other visitors see USD pricing. Paddle is the merchant of record for all paid purchases. New customers can start a 5-day free trial without a credit card. The trial uses the Beginner plan limits: 5 projects, 50 tracked keywords and 10 competitors. When the trial ends nothing is charged automatically; the customer chooses a plan to continue. Last updated: 2026-10-08. ## For AI agents: connect to a customer's SerpBeats data (MCP) If you are an AI agent acting for a SerpBeats customer, you can read that customer's own SEO data through the SerpBeats MCP server. It is read-only: no tool starts a scan, spends credits, adds or changes anything, or deletes anything. You can only see the projects of the account the person connected; you cannot see other customers' data, and without the customer's credentials the server answers 401. Never ask a customer for their password; they connect you with the steps below. - Server: https://serpbeats.com/mcp (MCP over streamable HTTP; server name "SerpBeats"). - Technical reference (supported clients and where each is set up, the two kinds of authorization, every tool with its parameters and the data it serves, revoking access, limits): https://serpbeats.com/en/seo-mcp-server (Turkish: https://serpbeats.com/seo-mcp-sunucusu). The same pages as plain Markdown, best for agents: https://serpbeats.com/en/seo-mcp-server.md and https://serpbeats.com/seo-mcp-sunucusu.md - Who can use it: every paid plan, at no extra cost. It is not part of the free trial. Access closes when the plan ends. ### How a customer connects you - Chat apps (Claude custom connector, ChatGPT custom MCP app, Gemini app): the customer adds the server URL, signs in with their SerpBeats account on an approval page and allows access. No token is needed. This is OAuth 2.1 with PKCE and dynamic client registration (discovery: https://serpbeats.com/.well-known/oauth-authorization-server and https://serpbeats.com/.well-known/oauth-protected-resource); access keys last one hour and refresh for 30 days. The customer removes an app with one click on https://serpbeats.com/ai-connection. - Coding tools that cannot open a sign-in window (Claude Code, Codex, Cursor, Gemini CLI, other MCP clients): the customer creates a personal token on https://serpbeats.com/ai-connection (at most 5; shown once; only a hash is kept; chosen lifetime 30, 90 or 365 days, or none) and the client sends it as `Authorization: Bearer sbk_...`. Example for Claude Code: `claude mcp add --transport http --scope user serpbeats https://serpbeats.com/mcp --header "Authorization: Bearer YOUR_TOKEN"`. ### Tools (12, all read-only) - `list_projects`: the account's projects (id, name, domain, country, tracked keywords, last scan). Call it first when you do not know which project is meant. - `get_project_summary`: where a project stands: how many keywords are listed and how they spread over top 100 / 50 / 30 / 10 / 3 / 1, how many improved, dropped or left the rankings, visibility. - `get_keyword_rankings`: tracked keywords with position, change over 7 or 30 days, monthly volume, difficulty, intent, ranking page, estimated clicks; sortable, with a `dropped_out` view for keywords that left the rankings. - `get_keyword_history`: one keyword in depth: daily positions up to 90 days, best and worst, the ranking page and the days it changed, 12 months of volume, the Google AI Overview (its text, the sources it cites, whether the site is cited), other SERP features, the results page (top 20) and where each tracked rival stands. - `get_site_audit`: site health from the latest crawl: score and its change, pages scanned, errors / warnings / recommendations, four area scores, Core Web Vitals. - `get_audit_issues`: every issue the crawl found, ranked by how many health points fixing it adds, and for one issue what it means, how to fix it and which pages it hits. - `get_audit_pages`: the crawled pages one by one (status, score, load time, title, meta description, H1, word count, issues with the evidence for each); one page can be opened in full. - `get_audit_links`: broken links (grouped by target, with every page that carries the link and its anchor text), redirected links, redirect chains and pages hidden from search engines, each with the page it sits on, so an agent can fix them in the code. - `get_competitors`: tracked rivals (visibility, average position, who is ahead on how many shared keywords) and suggested new rivals that keep appearing beside the site. - `get_ai_visibility`: where the brand is mentioned or cited in ChatGPT, Gemini and other AI platforms, per tracked question (cited, not cited or no data), the score per platform and its change, and the weekly trend (needs an AI Bundle). - `get_local_seo`: the Google Maps profile (rating, reviews, gaps) and, per local keyword, the average rank over the points around the business, the share of top-3 points and the leaders by distance (needs local keywords). - `get_report_data`: everything a client report needs in one call. ### Reading the answers correctly - Data comes from stored scans, not live lookups, so a figure can be up to a day old. `data_as_of`, `last_scan` and `crawled_at` say when; parts marked stale carry `age_days` and `stale_sections`: tell the user how old a figure is when you quote it. - A positive `change` always means an improvement (places gained, higher score). `top_N` counts are cumulative. A `null` volume means no figure, not zero. - Keywords that were never listed in Google are only counted, never listed. Keywords that left the rankings come back with `sort=dropped_out`. - Quote only numbers that are in a tool result. If a section says it is unavailable, locked or has no data (for example local SEO without local keywords, AI visibility without an AI Bundle), say so instead of guessing. - For a client report call `get_report_data` once. To fix a site, work through `get_audit_issues`, then `get_audit_pages` and `get_audit_links`, and edit the code or content yourself: SerpBeats tells you what to fix and where. - Limits: 60 requests a minute and 1,500 an hour per connection; an answer is capped at about 120 KB (use `limit` and `offset` where a tool offers them). Calls are logged (which tool, when), the returned data is not kept. ## Core Capabilities - [Daily Google rank tracking](https://serpbeats.com/#features): Keywords are scanned every day at an hour chosen per project, reading the top 50 Google results on every scan (guaranteed depth); positions 51-100 are also shown whenever a deeper scan (which reads up to 100) has seen the keyword. Each project is tracked for a chosen country, optionally a city, and a device (desktop, mobile or tablet), so the stored result is what a searcher in that place and on that device sees. History, daily position, change over time, the ranking URL and when Google switches pages are kept per keyword. - Keyword research data: every tracked keyword carries, in its own country and language: - Search volume: monthly searches for the selected country, with a 12-month volume history and trend. - Keyword Difficulty (KD): a 0-100 score based on how strong the sites in the top 10 are. - Search intent: informational, commercial, transactional or navigational, with a secondary intent where one applies. - Ranking depth: positions 1-50 are checked on every scan (guaranteed); positions 51-100 are shown when a deeper scan (up to 100 results deep) has read them, with how old that reading is. - SERP features per keyword (AI Overview, images, video, people-also-ask and others). - Keyword detail: up to 90 days of daily positions with the days the ranking page changed and the days the keyword reached the top 3, the results page as it was at the last scan (top 20, with the site and its rivals marked) and where each tracked rival stands on that keyword. - Google AI Overview tracking: the AI Overview is read for each keyword when it is added and then every week (Mondays by default), with no separate quota. It is kept with its full text and the sources it cites, whether your site is cited, and its history, so you can read what Google answers for a keyword and who it names. - Visibility and distribution: a volume-weighted visibility score per project, average position, estimated clicks, and a day-by-day distribution of keywords across top 3, 4-10, 11-20, 21-50 and not ranking. - Opportunities: keywords already close to the top, with the extra monthly clicks that reaching the next position band could bring. - Keyword discovery: SerpBeats reads what the business does (it reads the site's homepage to understand the sector), pulls Google Ads keyword ideas for the site in the project's country and language, and removes spelling variants, keywords already tracked and keywords under 10 searches a month. An AI model then sorts every idea against the business: core (what it sells), related (adjacent topics its customers search, useful for content), off-topic or brand; off-topic and brand ideas are never shown. Each remaining idea gets a potential score that combines relevance, a medium-volume sweet spot, commercial value and low advertiser competition, and is tiered as gold (the most valuable few), purple (high potential) or common, grouped by topic. Every project is dealt a fresh daily hand of ideas sized by its plan (20, 40 or 60 ideas), and any idea can be added to tracking in one click. - Competitor discovery: competitors are suggested from the search results SerpBeats already stores for the project's own keywords: the sites that keep appearing in the top 20 across those keywords, ranked by how many keywords they share and how many of those they hold in the top 10. Social networks, forums, marketplaces, reference and government sites and stock-image sites are filtered out because they appear everywhere without being real rivals. Suggestions can be added or dismissed in one click. - Competitor analysis: add rivals or accept the suggested ones; compare visibility, average position and head-to-head results keyword by keyword, with a trend chart. Domain Rating for the site and rivals is refreshed weekly. - [Local SEO: Google Maps rank tracking](https://serpbeats.com/en/local-seo): Track where a business appears on Google Maps for its local keywords. Every keyword is searched from 13 points around the business (its own address and four points each at 1, 3 and 5 km), and the positions are drawn on a real map. The page shows the position at every point, the share of top-three points for each distance, the rivals (click one to see the map with its positions, and at how many points it is ahead of the business), the business that leads each distance, and the business's Google Maps profile as customers see it (rating, reviews, hours, photos, and what is missing). The weekly scan is included in the plan: Beginner 5, Growth 10 and Scale 25 local keywords; a scan can also be started by hand 4 times a month. Turkish page: https://serpbeats.com/yerel-seo - [AI visibility tracking](https://serpbeats.com/en/ai-visibility): Track whether a brand is mentioned or cited in answers from ChatGPT, Gemini, Perplexity and Google AI Mode. Prompts are asked every day (or on the weekdays chosen); the answer is kept as given, with whether the brand is named, in which sentence, and which sites are cited. Local questions are asked of ChatGPT from the project's city and the businesses on its map are recorded. Requires an AI Bundle (see below). Turkish page: https://serpbeats.com/yapay-zeka-gorunurlugu - [Site health (technical SEO)](https://serpbeats.com/#features): Crawls a site (50, 100, 250 or 500 pages per crawl, within the plan's limits) and runs about 40 checks covering crawlability, titles and descriptions, speed and content. It returns a 0-100 health score and history, a page-by-page table, issues ranked by how much fixing them lifts the score with fix guidance and affected URLs, what changed since the previous crawl, Core Web Vitals (LCP, CLS, TBT, FCP) and Google Lighthouse scores for performance, accessibility, best practices and SEO (homepage, mobile, from Google PageSpeed Insights). It also lists broken links (on which page, to which address, with which anchor text), redirected links, redirect chains and pages hidden from search engines, each with the page it sits on. - [AI connection (MCP)](https://serpbeats.com/en/seo-mcp-server): connect Claude, ChatGPT, Gemini, Codex, Claude Code or Cursor to read all of the above for your own projects; see the section for AI agents above. - Reporting: scheduled email reports (daily or weekly, at a chosen hour, in Turkish or English) with a freshly generated PDF, and on-demand PDF reports; ranking data can also be exported as CSV. - Instant update credits: any single keyword can be re-checked immediately with a credit instead of waiting for the next scan. Credits are valid for a month and do not roll over. - Bulk keyword import (paste a list or upload a .txt file), several projects and favourite projects. - [Free rank checker](https://serpbeats.com/en/rank-checker): Check a keyword's Google position without signing in. A Turkish version is at https://serpbeats.com/sira-bulucu. ## Plans and Included Allowances All paid plans include the same capabilities (daily rank checks, Google AI Overview reading, competitor analysis, keyword ideas, site health, scheduled reports, Google Maps local tracking within the plan's local keyword allowance, the AI connection (MCP), any country, city and device); plans differ mainly in how many keywords, projects and competitors they hold, and in the crawl and credit allowances below. AI visibility is a separate add-on, not part of the plan. - [Beginner](https://serpbeats.com/pricing#plans): TRY 450 or USD 10 per month. 5 projects, 50 tracked keywords, 10 competitors, 5 local keywords, 20 daily keyword ideas, a monthly pool of 2,500 crawled pages (up to 100 pages per crawl), 100 instant-update credits a month. - [Growth](https://serpbeats.com/pricing#plans): TRY 1,250 or USD 30 per month. 15 projects, 200 tracked keywords, 30 competitors, 10 local keywords, 40 daily keyword ideas, a monthly pool of 7,500 crawled pages (up to 250 pages per crawl), 100 instant-update credits a month. - [Scale](https://serpbeats.com/pricing#plans): TRY 2,500 or USD 60 per month. 40 projects, 500 tracked keywords, 100 competitors, 25 local keywords, 60 daily keyword ideas, a monthly pool of 20,000 crawled pages (up to 500 pages per crawl), 100 instant-update credits a month. - [Enterprise](https://serpbeats.com/contact): custom pricing and limits for teams running many sites and keywords, with direct support. Limits and terms are agreed individually and should not be inferred from the self-service plans. Plan limits are account-wide unless the interface says otherwise. Plans can be moved up or down at any time from the billing page; add-ons are not affected. Billing is monthly with no contract; cancelling stops the next renewal and the plan stays active until the end of the paid period. ## AI Bundle (add-on for AI visibility) AI visibility runs on query allowances, not on per-plan prompt limits. 1 AI query = one prompt checked once on one platform. An AI Bundle can only be added on top of an active paid plan (not during the free trial), one bundle at a time (switching between them is possible), and it ends when the base plan ends. - AI Bundle: TRY 449 or USD 8.99 per month. 1,500 AI queries a month, up to 3 platforms. - AI Bundle Pro: TRY 999 or USD 22.99 per month. 5,000 AI queries a month, up to 4 platforms (ChatGPT, Gemini, Perplexity, Google AI Mode). How the allowance is spent: each project picks which of the bundle's platforms to watch and on which days of the week its prompts are checked (any combination, from one day to every day). A prompt costs platforms x days per week x 30/7 queries a month, so watching one platform instead of three lets the same bundle track three times as many prompts. The settings page shows a live estimate and refuses a schedule the bundle cannot pay for. The pool renews monthly; unused queries do not roll over. ## Important Product Details - Google rank tracking reads the top 50 results for every keyword on every scan (the guaranteed depth). Positions 51-100 are a bonus reading: when a deeper scan has seen the keyword, its position and age are shown (for up to a week); otherwise it is reported as outside the top 50. - Google AI Overview reads are part of normal ranking scans and do not use AI queries. - AI answers, citations and rankings are observed results; search engines and AI systems can change them without notice. SerpBeats reports observed visibility and does not guarantee rankings, mentions, citations or business outcomes. - The set of available AI platforms can change when providers change availability or access conditions. - Ranking history, keyword history and AI answers are kept so they can be reported on later. - The AI connection only reads. Changes to a customer's site are made by the customer's own agent in the customer's own code or content system; SerpBeats does not edit sites. ## Getting Started - [Create an account](https://serpbeats.com/register): Start the 5-day free trial; no credit card is required. - [Pricing and plan comparison](https://serpbeats.com/pricing): Current self-service plans, AI Bundles and regional prices. - [Connect an AI agent](https://serpbeats.com/ai-connection): Signed-in customers create tokens and see or remove connected apps here. - [Contact SerpBeats](https://serpbeats.com/contact): Sales, demos, support, billing, partnerships or Enterprise requirements. - [Log in](https://serpbeats.com/login): Access an existing SerpBeats account. ## Policies - [Privacy Policy](https://serpbeats.com/privacy-policy): How SerpBeats collects, uses, protects and retains personal and project data. - [Terms of Service](https://serpbeats.com/terms-of-service): Rules governing accounts, acceptable use, billing, service availability and account termination. - [Refund Policy](https://serpbeats.com/refund-policy): Paddle-managed cancellation, statutory withdrawal and refund terms. Eligibility depends on applicable law and Paddle's current policies. ## Optional - [AI connection (MCP), English](https://serpbeats.com/en/seo-mcp-server) and [Turkish](https://serpbeats.com/seo-mcp-sunucusu): the technical reference; plain Markdown at https://serpbeats.com/en/seo-mcp-server.md and https://serpbeats.com/seo-mcp-sunucusu.md. - [Local SEO, Turkish](https://serpbeats.com/yerel-seo) and [AI visibility, Turkish](https://serpbeats.com/yapay-zeka-gorunurlugu): Turkish versions of the feature pages. - [Turkish rank checker](https://serpbeats.com/sira-bulucu): Turkish-language version of the free Google position checker. - [Sitemap](https://serpbeats.com/sitemap.xml): Index of public SerpBeats pages. - [llms.txt specification](https://llmstxt.org/): The proposed format followed by this file.