--- title: "Rankscale MCP Server Docs | Rankscale" url: https://rankscale.ai/docs/mcp description: "How to connect Claude, ChatGPT, Cursor, and other MCP clients to Rankscale: server URL, permissions, tools, security, and troubleshooting." --- # Rankscale MCP Server Docs | Rankscale We now support Meta Muse Spark. [Track your Spark prompts](/features/ai-rank-tracker/meta-muse-spark) # Rankscale MCP server Connect Claude, ChatGPT, Cursor, or any Model Context Protocol client to your Rankscale data. Reading is always included. Creating content and spending credits are separate permissions that you turn on only if you want them. Server URL Available on Essentials, Pro, Pro+, Growth, and Enterprise plans. Nothing is installed on your machine. ## Connect ### Claude (web and desktop) - Open Settings, then Connectors, and choose Add custom connector. - Enter https://mcp.rankscale.ai/mcp and confirm. - Sign in to Rankscale in the tab that opens, pick the permissions you want, and choose Allow access. ### ChatGPT - Add a custom connector (app) with the URL https://mcp.rankscale.ai/mcp, then sign in to Rankscale when asked. ### Cursor and other MCP clients - Add a remote MCP server with the URL https://mcp.rankscale.ai/mcp. Any client that supports Streamable HTTP and OAuth can connect. Your client registers itself automatically (OAuth Dynamic Client Registration or Client ID Metadata Documents). There is no API key to copy. ## Permissions When you connect, Rankscale shows a consent screen. Read access is always granted. Each write permission is its own switch and starts off. ### Read your Rankscale data Always included. Lets your assistant read your brands, search terms, topics, executions, analytics, and saved dashboard views. Always on ### Create and update your search terms, topics, and dashboards Optional, off by default. New search terms are created inactive, so nothing runs or spends credits until you activate it. Off unless you turn it on ### Run search terms and activate tracking Optional, off by default. Spends your Rankscale credits. Every credit-spending call shows the cost first and waits for your confirmation. Off unless you turn it on Every tool that changes data is marked as a write, so Claude asks you before it runs one. Your assistant cannot delete brands or your account, and cannot change brand or competitor settings. ## Tools ### Read: brands, search terms, and runs Look up how your account is set up and what recent runs returned. Requires mcp:read. | Tool | What it does | | --- | --- | | rankscale_brands_list | Lists the brands you own or that are shared with you. | | rankscale_brands_get | Returns one brand by id. | | rankscale_search_terms_list | Lists the search terms tracked for a brand. | | rankscale_search_terms_get | Returns one search term and its configuration. | | rankscale_topics_list | Lists the topics that group a brand’s search terms. | | rankscale_executions_list | Lists recent runs as compact summaries (paginated). | | rankscale_executions_get | Returns one run in full: the AI answer, cited URLs, and mentioned companies. | | rankscale_operation_get | Reports the progress of a “run now” request. | ### Read: analytics Each analytics tool accepts the same filters (time frame, engine, topic, tags, query) and returns a summary plus an optional time series. Requires mcp:read. | Tool | What it does | | --- | --- | | rankscale_dashboard_summary | Your brand’s topline: AI Visibility Score, mentions, citations, sentiment, average position. | | rankscale_competitors_summary | Competitor leaderboard and per-competitor trend. | | rankscale_citations_summary | Top cited URLs and the sites that cite you most. | | rankscale_sources_summary | Domains that appear in AI source boxes next to you. | | rankscale_sentiment_summary | Sentiment breakdown and the keywords driving it. | | rankscale_shopping_summary | Products, providers, buybox winners, price and rating signals. | | rankscale_local_results_summary | Google local and Maps place results across AI engines: Share, average position, detection rate, top places, and prompts where the map results leave you out. | | rankscale_follow_up_forms_summary | ChatGPT follow-up form rate, top clarifying questions, and usual pre-selected answers. | | rankscale_custom_dashboard_views_list | Lists a brand’s saved dashboard views. | | rankscale_custom_dashboard_view_get | Returns one saved view’s full configuration. | ### Write: search terms, topics, and dashboard views Only available if you turn on the content permission when you connect. None of these tools spend credits. Requires mcp:write:content. | Tool | What it does | | --- | --- | | rankscale_search_terms_create | Creates a search term in the inactive state. Duplicate-safe. | | rankscale_search_terms_update | Updates a search term’s allowed fields (for example query, tags, cadence, engine). | | rankscale_search_terms_deactivate | Pauses 1 to 50 search terms. Shows a preview and asks you to confirm first. | | rankscale_topics_create | Creates a topic (a grouping for search terms). | | rankscale_topics_assign_terms | Groups search terms under a topic. Moving a term between topics asks you to confirm first. | | rankscale_topics_unassign_terms | Removes search terms from their topic. | | rankscale_custom_dashboard_view_create | Saves a new dashboard view. | | rankscale_custom_dashboard_view_update | Renames or reconfigures a saved view. | | rankscale_custom_dashboard_view_set_default | Makes one saved view the brand default. | ### Run tracking (spends credits) Only available if you turn on the credit-spending permission when you connect. Both tools show the cost first and wait for your confirmation. Requires mcp:execute. | Tool | What it does | | --- | --- | | rankscale_search_terms_activate | Starts recurring tracking on 1 to 50 inactive terms. Spends credits on every scheduled run. | | rankscale_search_terms_execute | Runs 1 to 20 terms once, right now, without changing their schedule. Spends credits. | ## Security, data, and revoking access - Tools only return data the signed-in Rankscale user can already see in the dashboard. Your assistant asks for one piece of data at a time during a conversation; it never gets a bulk export. - Sign-in uses OAuth 2.1 with PKCE. Access tokens last one hour and refresh automatically. Rankscale never issues long-lived API keys for MCP. - See and revoke every connected assistant at [Settings, Connections, MCP](https://rankscale.ai/dashboard/settings/connections/mcp). Revoking ends access for that connection immediately. - How we handle MCP data is described in our [privacy policy](/privacy#mcp). ## Troubleshooting and support Still stuck? Email [support@rankscale.ai](mailto:support@rankscale.ai). ## Reference ## For an optimal online experience We and our partners use technologies such as cookies to store and access information on your device to analyze site traffic, improve our service, and measure the performance of our marketing. You can accept all cookies, manage your preferences, or decline non-essential cookies. You can change your choice at any time from the Privacy Policy page. [Privacy Policy](https://rankscale.ai/privacy) ([md](https://rankscale.ai/privacy.md)) [Terms of Use](https://rankscale.ai/terms) ([md](https://rankscale.ai/terms.md)) [Imprint](https://rankscale.ai/imprint) ([md](https://rankscale.ai/imprint.md))