The Ranktracker MCP Server Is Now Live

Intro

The Ranktracker MCP server is officially live.

You can now connect supported AI assistants and development tools directly to your Ranktracker account, including ChatGPT, Claude, Cursor and other clients that support the Model Context Protocol.

Once connected, you can ask questions about your SEO performance in natural language.

For example:

• Which of my tracked keywords dropped the most this week?
• Show me the keywords currently ranking between positions 11 and 20.
• Which keywords have gained visibility during the last 30 days?
• Summarise the backlink trend for my website.
• What technical issues were discovered in my latest website audit?
• Which of my tracked keywords appear in an AI Overview?
• Show me the most recently discovered backlinks.
• What is the status of my latest SEO report?

The AI assistant selects the appropriate Ranktracker tool, securely retrieves the relevant account data and uses it to answer your question.

There is no need to manually export a CSV, switch between multiple reports or build a custom integration before you can start analysing your data.

MCP access is included with every Ranktracker plan.

Read the complete documentation at:

https://developers.ranktracker.com/mcp

What Is the Model Context Protocol?

The Model Context Protocol, commonly known as MCP, provides a standard way for AI applications to connect to external tools and data sources.

An AI model normally only has access to the information contained in its conversation and any general knowledge available to it. It cannot automatically see the projects, keywords, reports or audits inside your Ranktracker account.

An MCP connection changes that.

Once you authorise Ranktracker, the AI client can discover a defined collection of Ranktracker tools. When you ask a relevant question, the client can choose the correct tool, provide the required account identifiers and retrieve the data needed to answer.

Instead of manually navigating through several sections of Ranktracker, you can ask the question conversationally.

For example, when you ask:

Which keywords for example.com have lost the most positions this month?

The assistant can:

1. Find the relevant tracked domain.
2. Retrieve its tracked keywords.
3. inspect the ranking history.
4. Compare current and previous positions.
5. Summarise the most significant losses.

The Ranktracker MCP server handles the connection between the AI assistant and your account.

Connect Ranktracker to ChatGPT, Claude and Cursor

The Ranktracker MCP server is a hosted remote server that works with clients supporting Streamable HTTP MCP connections.

It can be connected to:

• ChatGPT
• Claude Desktop
• Claude Code
• Cursor
• MCP Inspector
• Other compatible MCP clients
• Internal AI agents and developer tools

The server is available through one hosted endpoint:

https://api.ranktracker.com/mcp

You do not need to download or host a separate Ranktracker server.

For most interactive AI tools, OAuth is the recommended connection method. You add the Ranktracker server URL to the client, approve access through your browser and the connection is created automatically.

No Custom Code Required

One of the biggest benefits of MCP is that you do not need to build the integration yourself.

With a traditional API integration, a developer usually needs to:

• Read the API documentation
• Create requests for individual endpoints
• Handle authentication
• Process the response
• Store or display the data
• Build an interface for the user

The Ranktracker MCP server presents clearly defined tools directly to the AI client.

The assistant discovers those tools automatically and decides which ones to use based on your request.

That means a user can connect Ranktracker and start asking questions without writing code or creating a separate reporting application.

Developers can still use the Ranktracker REST API for fully automated systems, scheduled data pipelines and custom applications. MCP is designed for cases where a person or AI agent is actively exploring the data.

What Ranktracker Data Can AI Assistants Access?

The initial Ranktracker MCP release provides 14 tools.

Thirteen tools are read-only. One additional tool can safely generate a fresh run of an existing report.

The available tools cover tracked websites, keywords, rankings, competitors, keyword research, backlinks, reports and website audits.

Tracked domains

Your assistant can list the domains available in your Ranktracker account and retrieve the configuration of a specific project.

This provides the starting point for most conversations because Ranktracker resources are connected using unique project identifiers.

Example questions include:

• List all the websites in my Ranktracker account.
• Find my project for example.com.
• Show me the tracking configuration for this domain.
• What monitoring limits are configured for this project?

The connection can only see domains belonging to the Ranktracker account that authorised it.

Tracked keywords and current rankings

Your assistant can retrieve the tracked keywords for a domain together with their latest ranking results.

This makes it possible to ask:

• What are my highest-ranking keywords?
• Which keywords currently rank on page one?
• Show me keywords ranking between positions 11 and 20.
• Which tracked keywords are outside the Top 100?
• Which keywords are currently paused?
• Show me the top 50 keywords for this domain.

The keyword data is ordered by the best organic position by default, allowing the assistant to identify the strongest current rankings quickly.

Historical ranking movements

The MCP server can retrieve daily ranking history for an individual tracked keyword.

This includes:

• Organic position
• Absolute SERP position
• Local-pack position
• SERP feature ownership
• AI Overview visibility
• Ranking history over a selected date range

The default date range covers the previous month, but the client can request specific start and end dates.

You can ask questions such as:

• How has this keyword moved during the last 30 days?
• When did this keyword enter the Top 10?
• Did we lose our featured snippet?
• Has our local-pack position improved?
• Is this domain included in the current AI Overview?
• What was the keyword’s best position this month?

This makes it easier to investigate ranking changes without manually opening and comparing individual charts.

Competitor data

The assistant can list the competitor websites configured for a tracked domain.

Example questions include:

• Which competitors are being monitored for this project?
• List the competitors attached to example.com.
• Are we monitoring competitor.com?
• Summarise the competitive set for this website.

The MCP server only retrieves competitors that have already been configured in the Ranktracker account. It does not automatically add new competitors or change project settings.

Keyword Finder research

The assistant can retrieve existing Keyword Finder research sessions from your account.

This allows you to find and discuss research that has already been completed.

You could ask:

• Show my latest keyword research sessions.
• Which keyword research projects are available?
• Find the research session created for our new product category.
• Summarise the keyword research available in my account.

The MCP connection does not begin new paid keyword research. It only reads existing research sessions, helping prevent an AI client from consuming allowances without your knowledge.

Backlink monitoring

The Ranktracker MCP tools can retrieve backlink information for a tracked domain.

Available backlink data includes:

• Source URL
• Target URL
• Anchor text
• Dofollow status
• Spam score
• Domain and page ranks
• First-seen date
• Last-seen date

This means you can ask:

• Show me the latest backlinks discovered for my website.
• Which new backlinks are dofollow?
• Are there any suspicious links with high spam scores?
• Which pages have received the newest links?
• What anchor text is being used most frequently?
• Show the strongest recently discovered backlinks.

Backlinks are returned with the newest links first, making the connection particularly useful for reviewing recent link-building activity.

Backlink profile trends

In addition to individual links, the assistant can retrieve historical backlink profile summaries.

These summaries can include:

• Domain rank
• Total backlinks
• New backlinks
• Lost backlinks
• Referring domains
• Historical profile snapshots

Example questions include:

• Summarise our backlink growth this month.
• Have we gained or lost referring domains?
• When did the largest backlink decline occur?
• Is our overall backlink profile growing?
• Compare the newest backlink snapshot with the previous one.

This allows an AI assistant to explain the trend rather than simply returning a list of links.

Monitoring summaries

The assistant can retrieve a combined monitoring summary for a domain covering areas such as:

• Keyword monitoring
• Backlink monitoring
• Website-audit monitoring

This can be used to request a concise overview of a website without checking every Ranktracker tool individually.

For example:

Give me an SEO monitoring summary for example.com.

The client can retrieve the available monitoring data and turn it into a clear list of changes, risks and areas requiring attention.

SEO reports

Your assistant can list the reports already configured in your Ranktracker account.

It can also retrieve information about a particular report, including:

• Report schedule
• Recipients
• Latest run
• Recent run history
• Current report status
• Completed report file URL

You can ask:

• List my configured SEO reports.
• When was the latest client report generated?
• Did this week’s report complete successfully?
• Show me the latest available report.
• Which recipients are configured for this report?
• What is the current report schedule?

This can be particularly useful for agencies managing multiple recurring client reports.

Run an existing report

The MCP server includes one carefully controlled write action: run_report.

This lets an AI assistant generate a fresh run of a report that you have already configured in Ranktracker.

The action is intentionally limited.

It:

• Does not create a new report
• Does not email report recipients
• Does not purchase additional data
• Does not add keywords
• Does not change tracked projects
• Does not consume keyword-tracking credits
• Cannot repeatedly run the same report within one reporting period

The generated report appears inside Ranktracker, and the tool can return its status and file when the run is completed.

This gives users a useful action without allowing an AI client to make broad or expensive account changes.

Website audits

The assistant can list the website audits attached to a domain and retrieve the results of a specific audit.

Available audit information includes:

• Audit status
• Crawl progress
• Crawl statistics
• Audit scores
• Warnings
• Technical SEO checks
• SSL status
• HTTP/2 status
• Sitemap checks
• robots.txt checks
• Indexing information

You could ask:

• What were the main problems in our latest website audit?
• Did the audit identify any indexing issues?
• Is the website using HTTP/2?
• Were there any sitemap or robots.txt problems?
• Summarise the highest-priority technical warnings.
• Has the latest audit finished?
• Compare the number of warnings with the previous audit.

This gives SEO teams a faster way to understand technical audit results and turn them into prioritised actions.

Ask Questions Across Different Ranktracker Tools

The real value of an AI connection is not limited to retrieving one field at a time.

An assistant can call several tools during the same conversation and combine the results into a more complete answer.

For example, you could ask:

Prepare a weekly SEO summary for example.com.

Depending on the request and available information, the assistant could review:

• Current keyword positions
• Recent ranking movements
• AI Overview visibility
• New and lost backlinks
• Backlink profile trends
• The latest website-audit status
• Configured competitors
• Report availability

It can then present the findings as a management summary, a client update, a list of priorities or a technical action plan.

You can continue the conversation with follow-up questions such as:

• Which ranking loss should we investigate first?
• Show me the supporting keywords.
• Turn this into a client-friendly summary.
• Separate quick wins from longer-term issues.
• Which technical warning could have the largest SEO impact?

Instead of repeatedly exporting and reformatting data, the analysis can happen within the conversation.

How to Connect Ranktracker With OAuth

OAuth is the recommended option for ChatGPT, Claude, Cursor and other interactive clients.

The general setup process is:

1. Open the MCP or connector settings in your AI client.
2. Add the Ranktracker MCP endpoint:

https://api.ranktracker.com/mcp 3. Leave the manual authentication field empty. 4. Allow the client to discover the Ranktracker OAuth flow. 5. Sign in to Ranktracker in the browser window. 6. Review the requested permissions. 7. Approve the connection. 8. Return to the AI client and confirm that the Ranktracker tools are available.

Only a Ranktracker account administrator can approve a connection because the connected client can access data across the authorised account.

Connecting ChatGPT

In ChatGPT:

1. Open Settings.
2. Open Connectors.
3. Select Add custom connector.
4. Paste the Ranktracker MCP server URL.
5. Complete the browser authorisation process.

No manual authentication header is required when using OAuth.

Connecting Claude Code

Claude Code users can add Ranktracker with:

The browser consent screen opens when authentication is required.

Connecting Claude Desktop or Cursor

Claude Desktop and Cursor can use the following server configuration:

After saving the configuration, connect the server and approve access through the Ranktracker browser consent screen.

The exact location of the configuration file depends on the client. Complete client-specific instructions are available in the Ranktracker MCP documentation.

MCP Permissions and Account Security

The OAuth consent screen shows which application is requesting access, which Ranktracker account will be connected and which permissions are being requested.

The available permission scopes are:

• mcp:read — read projects, keywords, rankings, competitors, backlinks, audits and reports
• mcp:write — generate a new run of an existing report

OAuth access tokens last for one hour and are refreshed automatically by the connected client. The connection remains active until it is revoked.

The client can only access information belonging to the Ranktracker account that approved the connection.

Most available MCP actions are read-only. Actions that could consume paid credits or make major tracking changes are deliberately excluded.

An AI client cannot use the MCP server to:

• Add new tracked domains
• Add or remove keywords
• Purchase additional data
• Start new Keyword Finder research
• Refresh keyword rankings
• Change tracking frequencies
• Delete projects
• Modify competitors
• Email reports to recipients

These restrictions allow users to analyse their SEO information without handing an AI assistant unrestricted control over their Ranktracker account.

OAuth or MCP Token?

Most users should connect with OAuth.

OAuth does not require you to copy, save or paste a permanent token. The AI client opens Ranktracker in your browser, asks for approval and manages the temporary access credentials automatically.

A dedicated MCP token is also available for:

• CI pipelines
• Scripts
• Internal agents
• Automated environments
• Clients that do not support OAuth

MCP tokens begin with:

tkn_mcp_usr_...

They are separate from normal Ranktracker REST API keys and only work with the MCP endpoint.

A REST API key beginning with tkn_usr_... will not work as an MCP credential. Similarly, an MCP token cannot be used with Ranktracker REST API endpoints.

Keeping the credentials separate means an MCP token can be revoked without affecting an organisation’s existing API integrations.

Ranktracker MCP vs the Ranktracker REST API

The MCP server and REST API connect to the same Ranktracker account data, but they are designed for different workflows.

Ranktracker MCP	Ranktracker REST API
Designed for AI assistants and agents	Designed for applications and backend systems
Supports conversational data exploration	Uses fixed programmatic endpoints
The AI selects the appropriate tool	A developer selects and calls each endpoint
Primarily read-only	Supports broader create, read, update and delete operations
Best when a person or agent is asking questions	Best for scheduled and automated data movement
Uses OAuth or a dedicated MCP token	Uses a REST API key

Use MCP when you want to explore, explain or summarise your SEO data through an AI conversation.

Use the REST API when you need to build a dashboard, maintain a data warehouse, synchronise systems or run a scheduled integration.

Many teams will use both: MCP for conversational analysis and the REST API for structured automation.

Built to Handle Large SEO Accounts

Keyword and backlink tables can contain a significant amount of data.

By default, Ranktracker MCP returns structured, readable JSON. For larger result sets, users can enable the optional TOON response format by adding:

?format=toon

to the MCP endpoint.

TOON returns the same underlying information in a more compact tabular format. According to the Ranktracker documentation, this can reduce token usage by approximately 40% to 60% for large datasets.

The standard endpoint is:

https://api.ranktracker.com/mcp

The token-optimised endpoint is:

https://api.ranktracker.com/mcp?format=toon

Most users can leave the connection on the standard JSON format. The compact option is primarily useful for larger keyword and backlink datasets.

Rate Limits

The Ranktracker MCP endpoint has separate limits from the REST API:

• 200 requests per minute for each MCP credential
• 300 requests per minute for each IP address

Requests exceeding the limit return an HTTP 429 response and should be retried after a delay.

These limits are sufficient for normal conversations while helping protect accounts and infrastructure from excessive automated requests.

MCP Access Is Included With Every Ranktracker Plan

The Ranktracker MCP server is available across every Ranktracker subscription plan.

This means customers on Launch, Growth, Scale and Authority can connect supported AI clients to the data inside their account.

The amount of data available will depend on the projects, keywords, backlinks, audits and reports already available under the account’s existing subscription and allowances.

There is no separate MCP subscription required to begin connecting your Ranktracker account to a supported AI client.

Start Using Ranktracker With Your AI Assistant

The Ranktracker MCP server creates a new way to work with SEO data.

Instead of moving between individual screens, downloading spreadsheets and manually preparing summaries, you can connect your AI assistant and ask direct questions about your account.

You can analyse:

• Tracked keywords
• Ranking movements
• AI Overview visibility
• Competitors
• Keyword research
• New and lost backlinks
• Backlink profile trends
• Website audits
• Scheduled reports

The connection is quick to set up, works with leading MCP clients and is designed with controlled permissions that prevent costly or destructive account changes.

Connect your preferred AI client and explore the complete setup instructions at:

https://developers.ranktracker.com/mcp