MCP Server
Kvery ships a built-in MCP (Model Context Protocol) server, so AI assistants that speak MCP can discover and run your published queries as tools — letting an assistant answer questions straight from your databases through Kvery.
What MCP exposes
Through the MCP server an assistant can:
- List queries available to it.
- List groups.
- Search queries by name or content.
- Get a query's details.
- Execute a query (with parameters) and read the result.
- Read the tags associated with a token's access.
Access is always scoped by the token you issue (see below) — an assistant only ever sees what its token is allowed to see.
Connecting
The server supports two transports plus a REST surface:
| Transport | Endpoint | Use |
|---|---|---|
| JSON-RPC | POST /mcp | standard MCP protocol calls |
| SSE | GET /mcp | streaming MCP clients |
| REST | GET /mcp/queries, /mcp/groups, /mcp/search, GET/POST /mcp/queries/{id}… | direct integration |
Point your MCP-capable client at the /mcp endpoint and authenticate with a Kvery
MCP token.
MCP tokens
You manage access with MCP tokens, which you create and revoke from Kvery:
- Every token is prefixed with
kvery_mcp_, so it's easy to recognise. - Tokens expire after 90 days by default.
- Each token carries its own scope (the queries and tags it may reach), so you can hand different assistants different, least-privilege access.
- Revoke a token at any time to immediately cut off access.
An MCP token grants an assistant the ability to run your queries. Store it like a password, scope it tightly, and revoke tokens you no longer use.
Tips
- Publish and name queries clearly so an assistant can find the right one via search.
- Use tags to group the queries a token should expose.
- Combine MCP with rate limits to keep automated usage in check.