Rankry MCP Server

The Rankry MCP server lets AI clients read your AI visibility data directly. Connect Claude, ChatGPT, Cursor, n8n, or any tool that speaks the Model Context Protocol, and it can answer questions about your Score, Visibility, Sentiment, competitors, cited Sources, and Google Search Console performance using live data from your account. The server is read-only and scoped to your own brands, and the numbers it returns are identical to the ones in your Rankry Dashboard.

What this page covers

What the MCP server is and what it can do
Creating and managing an API key
Connecting your MCP client
The 14 tools the server exposes
Rate limits and access requirements

What the MCP server is

MCP is an open standard for connecting AI assistants to external tools and data. Rankry runs an MCP server so an assistant can pull your visibility data into a conversation instead of you copying numbers out of the Dashboard. You can ask an assistant questions like “how is my brand doing on AI search this month?” or “which sources cite my competitors but not me?” and it will call the right Rankry tool and answer from your live report.

Two properties matter:

Read-only. Every tool only reads data. Nothing an assistant does through MCP can change your account, brands, prompts, or reports.
Private to your account. Each API key is tied to one user. Tools can only read the brands that key’s owner owns. There is no way to reach another account’s data.

Before you start

You need an active Rankry subscription and an API key. Creating keys is free, but calling the MCP server requires a current subscription. If your subscription lapses, existing keys stop working until you reactivate.

Create an API key

Open Account Settings and find the API & MCP card.
Select Create API key and give it a name that tells you where it will be used, for example “Claude Desktop”.
Copy the key immediately. The full key is shown only once, right after you create it. It starts with rnk_.
Store it somewhere safe. If you lose it, revoke it and create a new one.

You can hold up to 20 active keys per account. Each key in the list shows its name, a masked preview, when it was created, and when it was last used. To retire a key, select the trash icon next to it and confirm. Revoking a key takes effect immediately.

Connect your MCP client

The server speaks Streamable HTTP at a single endpoint, authenticated with your API key as a bearer token.

Endpoint: POST https://rankry.ai/app/mcp
Authentication: Authorization: Bearer rnk_your_key_here

Most MCP clients accept a small JSON configuration. Use the HTTP transport, the endpoint above, and your key in the Authorization header:

{
  "mcpServers": {
    "rankry": {
      "type": "http",
      "url": "https://rankry.ai/app/mcp",
      "headers": { "Authorization": "Bearer rnk_your_key_here" }
    }
  }
}

This works with Claude, ChatGPT, Cursor, n8n, and the OpenAI and Gemini SDKs, along with any other client that supports the Model Context Protocol. The exact place you paste this configuration depends on the client, but the values are always the same: the endpoint and your bearer key. The API & MCP card in Settings also shows your server URL and an example configuration you can copy.

Available tools

The server exposes 14 tools. Most take an optional projectId (the brand id from list_projects); if you omit it, the tool uses your most recently updated brand. Where a tool accepts an engine, the options are chatgpt, claude, gemini, perplexity, and grok. Where it accepts a period, the options are 7d, 30d, 90d, and 365d.

Brands and reports

Tool	What it returns
`list_projects`	Every brand in your account with its current Score and last report date. Call this first when you track more than one brand.
`list_reports`	A brand’s report history, newest first, with score, status, and dates.

Visibility and performance

Tool	What it returns
`get_brand_summary`	The headline view: overall Score and its change, per-engine visibility, mention and citation rate, sentiment, cross-model consistency, and top insights. Pass a `period` to include a trend.
`get_visibility`	Per-engine visibility detail: weighted visibility, mention rate, prompts analyzed, and which content categories are strong, need work, weak, or invisible.
`get_sentiment`	Per-engine sentiment: the positive, neutral, and negative split, top positive themes, top concerns, and the single biggest deal-breaker AI models raise.
`get_competitors`	A leaderboard of your tracked competitors ranked by AI visibility, with each player’s visibility, average position, and sentiment, plus where you rank.
`get_insights`	The actionable insights from your latest report, each tagged with a severity and a status.
`get_history`	A time series of your Score and per-engine visibility, sentiment, position, and consistency over a period.
`get_top_prompts`	The prompts Rankry tested, ranked by how well your brand performs, with the engines that mentioned it, average rank, sentiment, and competitors named alongside it.
`list_sources`	The web sources AI engines cite in your space, aggregated by domain and ranked by citations. Each domain is typed and flagged as yours, a gap, or mixed, so you can find where competitors are cited and you are not.

Google Search Console

These tools require Google Search Console to be connected for the brand. Use get_integrations first to check.

Tool	What it returns
`get_integrations`	The integrations available on your plan and their connection status for a brand, including whether Search Console is connected.
`get_gsc_performance`	Search Console performance for your site: clicks, impressions, CTR, and average position, grouped by query, page, country, device, date, or hour.
`get_gsc_index_status`	URL inspection for a specific page: whether it is indexed, its coverage state, robots verdict, last crawl time, and which sitemaps contain it.
`get_gsc_sitemaps`	The sitemaps submitted to Search Console for your site, with download status, error and warning counts, and last-downloaded time.

Rate limits and access

The MCP endpoint is rate limited to protect the service. Each API key can make up to 60 requests per minute. If you exceed the limit, the server returns a 429 response with a Retry-After header telling you how many seconds to wait before retrying. Normal assistant use stays well under this.

Access requires an active subscription. Search Console tools additionally require Google Search Console to be connected on your plan. All data is read-only and scoped to the account that owns the key.

Frequently asked questions

Does Rankry have a public REST API?

Not yet. The MCP server is the supported way to access your data programmatically today. It gives read access to the same numbers you see in the Dashboard.

Do I need a paid plan?

Yes. Creating and revoking keys is free, but calling the MCP server requires an active subscription. Keys stop working if the subscription lapses and resume when you reactivate.

Can the MCP server change my data?

No. Every tool is read-only. An assistant connected through MCP can read your reports but cannot edit brands, prompts, or anything else in your account.

Can it see other people’s brands?

No. Each key is tied to one account, and tools can only read that account’s brands.

Which clients can I use?

Any client that supports the Model Context Protocol, including Claude, ChatGPT, Cursor, n8n, and the OpenAI and Gemini SDKs.

What if a tool says Search Console is not connected?

The Search Console tools need Google Search Console connected for that brand. Call get_integrations to check the status, then connect Search Console in your integrations settings.

How do I revoke access?

Open the API & MCP card in Account Settings and revoke the key. It stops working immediately. Revoke and recreate a key any time you think it has been exposed.