Tools Reference
Complete reference for all MCP server tools.
The IndexMind MCP server exposes 15 tools that your IDE can invoke. Tools are split into two categories: read-only tools that fetch data, and write tools that modify project state.
When the server is started with the --readonly flag, write tools are not registered and will not appear in your IDE's tool list. This is useful for shared environments or when you want to prevent accidental mutations.
Read-only tools
These tools are always available regardless of the --readonly flag.
| Tool | Description |
|---|---|
list_projects | List all projects accessible to the authenticated API key. Returns project IDs, names, and domains. |
get_project_context | Fetch full details for a specific project, including business context, keywords, and configuration. |
list_recommendations | List recommendations for a project, filterable by status (OPEN, IN_PROGRESS, DONE). |
get_recommendation_detail | Retrieve a single recommendation with full fix guidance, code samples, and impact assessment. |
list_visibility_queries | List AI visibility queries and their results for a project. Shows how AI models respond to target keywords. |
get_visibility_fix | Generate a copy-ready fix for a specific visibility issue. Returns structured content you can apply directly. |
list_sentiment_analyses | List sentiment analysis results for a project. Includes tone classification and framing details. |
get_sentiment_fix | Generate or retrieve actionable suggestions to improve sentiment for a specific analysis result. |
list_narrative_actions | List competitor narrative actions for a project. Shows identified narrative gaps and recommended responses. |
list_agent_actions | List pending and recent agent actions. Returns actions awaiting approval and recently executed actions. |
Write tools
These tools modify project data and are disabled when the server runs with --readonly.
| Tool | Description |
|---|---|
update_recommendation_status | Update the status of a recommendation. Accepts OPEN, IN_PROGRESS, or DONE. |
generate_narrative_actions | Create new narrative fix actions based on the latest competitor analysis data. |
update_narrative_action | Update the status or ownership of an existing narrative action. |
approve_agent_action | Approve a pending agent action, allowing it to execute. |
Write tools perform real mutations against your IndexMind project. Use the --readonly flag in environments where you want to prevent accidental changes, such as shared team servers or demo setups.
Resource URIs
In addition to tools, the MCP server exposes resources using the indexmind:// URI scheme. Resources provide structured data that your IDE can reference as context without invoking a tool.
Examples of resource URIs:
indexmind://projects
indexmind://projects/{project_id}
indexmind://projects/{project_id}/recommendations
indexmind://projects/{project_id}/visibility
indexmind://projects/{project_id}/sentiment
Your IDE may display these resources in a resource browser or allow you to attach them as context to a conversation. The exact behavior depends on your IDE's MCP implementation.
Tool invocation
Tools are invoked automatically by your IDE's AI assistant when it determines they are relevant to your prompt. You can also reference tools explicitly in your prompts:
- "List all open recommendations for this project"
- "Show me the visibility data for my target keywords"
- "Mark recommendation rec_456 as done"
- "Approve the pending agent action act_789"
The MCP server handles authentication, pagination, and error formatting so your IDE receives clean, structured responses.