MCP integration

The Model Context Protocol (MCP) lets AI agents use external tools, such as file systems, databases, browsers, and custom services. Taichu uses MCP servers on your own computer through the GWC for VS Code setting gwc.mcpServers. GWC Bridge separately connects supported AI apps to the computer and folders you choose to share.

How it works

  1. Add an MCP server — configure gwc.mcpServers in GWC for VS Code settings on the computer where Taichu runs.
  2. Your agent finds tools — Taichu discovers tools from the local MCP servers you configured for its agent.
  3. Your agent uses tools — Taichu calls a configured server from your computer and receives its structured tool result.
  4. Where tools run — local MCP server processes and their tool calls run on your computer. GWC Bridge connects external AI apps to the computer under the permissions you grant.

Connect your own MCP servers

In GWC for VS Code, configure your own MCP servers using thegwc.mcpServers setting. Taichu runs those servers locally. GWC Bridge is for connecting AI apps to your shared computer; there is no supported public cloud MCP registration API.

Taichu MCP →

Use Taichu's local file and search tools from an MCP-compatible AI client. Find setup steps and limits.

Custom MCP Server

Use a custom MCP tool server from Taichu on your own computer.

Custom MCP server

Add your own MCP-compatible tool server in the Taichu settings for GWC for VS Code:

// GWC for VS Code settings.json: local MCP servers
{
  "gwc.mcpServers": {
    "my-tool": {
      "command": "node",
      "args": ["./my-mcp-server.js"],
      "env": {
        "API_KEY": "${env:MY_API_KEY}"
      }
    }
  }
}