Gemini Docs

MCP server for retrieving relevant documentation from a knowledge base
  • typescript

14

GitHub Stars

typescript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "m-gonzalo-cosa-sai": {
      "command": "bun",
      "args": [
        "--watch",
        "path/to/repo/cosa-sai-mcp/src/index.ts",
        "--verbose"
      ],
      "env": {
        "GEMINI_API_KEY": "<your_gemini_api_key>"
      }
    }
  }
}

Gemini Docs MCP Server lets you access a curated Gemini-based documentation knowledge base directly through an MCP client. It uses the Gemini API to answer questions about documentation, supports multiple querying tools, and is designed to work well in environments like Roo/Cline by avoiding noisy web results and enabling richer, context-aware responses.

How to use

You interact with the Gemini Docs MCP Server through your MCP client. Enable the server in your client’s MCP settings, then issue questions that reference documentation or ask for guidance about how to use technologies covered in the knowledge base. Use the available tools to check feasibility, get hints, validate practices, or explore how-to examples. If you enable verbose mode, you’ll get additional debugging output to help you troubleshoot requests.

How to install

Prerequisites: you need Bun installed to run the MCP server locally.

Step 1: Install the MCP client via Smithery. This will set up the server configuration for your MCP client and ensure the server is managed by the client.

Step 2: Run the MCP server configuration snippet shown below in your client settings to start the server with verbose logging and the Gemini API key.

{
  "command": "bun",
  "args": [
    "--watch",
    "path/to/repo/cosa-sai-mcp/src/index.ts",
    "--verbose"
  ],
  "env": {
    "GEMINI_API_KEY": "<your_gemini_api_key>"
  },
  "disabled": false,
  "alwaysAllow": [
    "can_x_be_done",
    "hints_for_problem",
    "is_this_good_practice",
    "how_to_do_x"
  ],
  "timeout": 60
}

Additional notes

The server relies on a locally running Bun environment and expects you to provide a Gemini API key via the GEMINI_API_KEY environment variable. The configuration shown enables four tools that you can use in your MCP client: can_x_be_done, hints_for_problem, is_this_good_practice, and how_to_do_x. If you need debugging output, enable the verbose flag as shown.

Configuration and environment

The MCP server is started with a stdio configuration that runs Bun to watch the MCP entry point and expose verbose logging. The required environment variable is GEMINI_API_KEY, which should contain your Gemini access key.

Security and access control

Limit access to the MCP server by configuring your client to only allow the allowed tools and by keeping the GEMINI_API_KEY secret. Do not expose the API key in shared settings or logs.

Troubleshooting

If requests fail or you don’t see verbose logs, verify that Bun is installed, the GEMINI_API_KEY is correctly set, and the path to the MCP source index.ts is correct in the configuration. Ensure the MCP client is restarted after updating the server configuration.

Available tools

can_x_be_done

Check if a specific task can be done in a given technology.

hints_for_problem

Get hints for solving a specific problem.

is_this_good_practice

Check if a code snippet follows good practices.

how_to_do_x

Get examples and alternative approaches for a specific task.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational