Metabase

Metabase MCP Server

Metabase includes a built-in Model Context Protocol (MCP) server that lets AI clients connect directly to a Metabase instance. It uses the Streamable HTTP transport and builds on Metabase's Agent API to expose tools for searching, navigating, querying, visualizing, and creating/updating content - all scoped to the connecting user's permissions.

Endpoint

The MCP server is available at:

https://{your-metabase.example.com}/api/metabase-mcp

The legacy /api/mcp path still works as an alias for existing clients, but /api/metabase-mcp is the canonical URL to advertise.

Connecting a client

Point any MCP-compatible client at the /api/metabase-mcp endpoint. For example, with Claude Code:

claude mcp add metabase https://{your-metabase.example.com}/api/metabase-mcp --transport streamable-http

For Claude Desktop, create a custom connector using the same URL.

For Cursor, open Settings > MCP and add a new server with the type set to streamable-http and the URL:

https://{your-metabase.example.com}/api/metabase-mcp

Authentication

MCP clients authenticate via OAuth 2.0. Metabase runs its own embedded OAuth server - no external provider is needed.

The flow for a first-time connection:

• The client discovers Metabase's OAuth endpoints.

• The client registers itself with Metabase.

• The user is redirected to Metabase to log in and approve the connection.

• The client receives an access token scoped to the user's Metabase permissions.

Browser-based sessions (cookie auth) are also supported and receive unrestricted scopes.

Scopes

Access tokens are scoped to limit what tools a client can use:

Scope Grants access to agent:search search agent:resource:read read_resource (always granted to any authenticated caller; per-URI perm checks happen inside the dispatcher) agent:query:construct construct_query agent:query query agent:query:execute execute_query agent:sql:construct construct_native_query agent:sql:execute execute_sql agent:question:create create_question agent:question:update update_question (also covers "move card to collection") agent:question:execute execute_question agent:dashboard:create create_dashboard agent:dashboard:update update_dashboard agent:collection:create create_collection

Wildcard patterns (e.g. agent:*) match any scope with that prefix.

OAuth protected resource metadata is available at:

/.well-known/oauth-protected-resource/api/metabase-mcp

By default our consent screen grants access to all scopes without the opportunity to customize.

Available tools

The MCP server exposes these tools, dynamically generated from the Agent API endpoint metadata:

Discovery + read

Tool Description search Search for tables, metrics, cards, dashboards, and collections using keyword or natural-language queries. read_resource Read one or more Metabase entities by metabase:// URI. Covers database/schema/table/collection/question/dashboard/metric/transform navigation. Up to 5 URIs per call.

Query construction + execution

Tool Description construct_query Construct a query against a table or metric. Accepts the user's original prompt when available. Returns an opaque query_handle for use with execute_query or visualize_query. construct_native_query Construct a native (raw SQL) query for a database. Returns an opaque query_handle to feed create_question and save it. Does not execute the SQL; native handles are rejected by execute_query/query (use execute_sql to run raw SQL). query Query a table or metric directly. Supports pagination via continuation tokens. execute_query Execute a previously constructed query and return results with column metadata. execute_sql Execute a raw SQL query against a database. Requires the user to have native-query permission on the target database. Can be disabled instance-wide via the mcp-execute-sql-enabled setting. execute_question Run a saved question by id and return its rows + column metadata. Runs under the caller's permissions. Parameterized questions are not supported (returns an error).

Write

Tool Description create_question Save a query as a named question (card). Accepts a query_handle from construct_query (MBQL) or construct_native_query (native SQL). Saving native requires native-query DB permission. update_question Update a saved question. Patch semantics. Setting collection_id moves the card. Setting archived archives it. Replacing the query accepts a construct_query or construct_native_query handle. create_dashboard Create a new dashboard, optionally populated with saved questions (auto-positioned on the grid). update_dashboard Update a dashboard's metadata (name, description, collection, archived). create_collection Create a new collection. Optionally nested under a parent_collection_id.

Query results are limited to 200 rows per request. When more rows are available, the response includes a continuation_token that can be passed back to fetch the next page.

read_resource list responses cap at 25 items with truncated / total signals; drill into specific URIs to see more, or refine via search.

Resources

The server exposes MCP resources so clients can fetch supplementary content by URI without inflating tool descriptions.

Resource URI Description metabase://docs/construct-query.md Program syntax for construct_query and query: sources, operations, operator forms, worked examples, pitfalls.

The read_resource tool (above) uses a separate URI scheme to navigate Metabase entities (metabase://question/{id}, metabase://database/{id}/tables, etc.). The two URI namespaces are independent: metabase://docs/... is for static reference content fetched via MCP resources/read, while metabase://table/... and friends are entity URIs passed to the read_resource tool.

Supported JSON-RPC methods

Method Description initialize Initialize the MCP connection. Returns server capabilities and a session ID. notifications/initialized Client notification that initialization is complete. tools/list List available tools (filtered by the token's scopes). tools/call Call a tool with arguments. resources/list List available resources (filtered by the token's scopes). resources/read Read a resource by URI. Requires an initialized session. ping Keepalive ping.

Requests can be sent individually or as a JSON-RPC batch. The server responds with JSON or SSE depending on the Accept header.

Architecture

The implementation lives in these files:

api.clj - The HTTP handler. Parses JSON-RPC requests, validates authentication and session headers, enforces origin checks (DNS rebinding protection), and dispatches to the appropriate method. Supports both JSON and SSE response formats.

tools.clj - Tool dispatch and manifest generation. Builds the tool list from Agent API endpoint metadata, checks scopes, and routes tool calls through synthetic Agent API requests.

resources.clj - MCP resource registry and handlers. Holds documentation resources (like the construct_query reference) keyed by URI, with scope-based access control on resources/list and resources/read.

scope.clj - Scope matching logic. Supports exact matches, wildcard patterns, and the ::unrestricted sentinel for session-based auth.

Request flow

MCP client
 -> POST /api/metabase-mcp (JSON-RPC)
 -> Origin + session validation
 -> Auth: OAuth bearer token or browser session
 -> Scope check against requested tool
 -> Synthetic request to Agent API endpoint
 -> Response materialized as MCP content
 -> JSON or SSE back to client

Further reading

• MCP user docs

• Agent API source

• Model Context Protocol specification