KWDB

The KWDB MCP Server lets you interact with a KWDB database via the MCP protocol. You can run it locally or expose it over HTTP to support read and write operations, query planning, and consistent API responses for tools and MCP clients. This guide walks you through using the server, installing it, and understanding its configuration and security considerations.

How to use

Connect an MCP client to the KWDB MCP Server to run read and write queries, manage database objects, and view metadata. You can run the server locally in stdio mode for development or in HTTP mode for production-grade access. The server automatically limits large results for safety and provides structured error handling for tools and resources.

Two primary ways to access the MCP server are available: HTTP mode for remote clients and StdIO mode for local development. In HTTP mode, the server exposes an endpoint you can query at /mcp, while StdIO mode runs as a local process that you connect to directly.

Use these example start commands to run the server in the supported modes. The HTTP mode is the recommended option for production. The StdIO mode is convenient for development and testing on a local machine.

# HTTP mode (recommended)
CONNECTION_STRING="postgresql://<username>:<password>@<hostname>:<port>/<database_name>?sslmode=disable" PORT=8080 make run-http

# StdIO mode (local development)
./bin/kwdb-mcp-server "postgresql://<username>:<password>@<hostname>:<port>/<database_name>?sslmode=disable"

What you can do with the server

Read data with SELECT, SHOW, and EXPLAIN statements. Write data and modify the schema with INSERT, UPDATE, DELETE, and DDL commands like CREATE, ALTER, and DROP.

Available tools

read_query

Executes read-only statements such as SELECT, SHOW, and EXPLAIN to read data and return results. Automatically applies LIMIT 20 to SELECT queries without an explicit LIMIT to prevent large result sets.

write_query

Executes data modification and DDL statements including INSERT, UPDATE, DELETE, CREATE, ALTER, and DROP to modify data and schemas.