Firebase Docs MCP Server Setup

What is Firebase Docs MCP Server Setup?

Firebase Docs MCP Server Setup is a toolchain that indexes Firebase documentation into a local SQLite database and serves the indexed content through a Model Context Protocol (MCP) server using stdio transport. It is designed for developers who want to query Firebase documentation via MCP-compatible AI assistants or tools. The project includes a Go-based indexer, a Node.js MCP server, and a Genkit-based MCP client for testing.

How to use Firebase Docs MCP Server Setup?

1. Set the Gemini API key: export genaikey="APIKEY".
2. From the docs-mcp folder, run go run . to index Firebase docs into $HOME/.indexResp/db.sqlite.
3. Copy the database to the docs-mcp-server folder: cp $HOME/.indexResp/db.sqlite ..
4. Install dependencies and build: npm ci && npm run build.
5. Test the MCP server: npx @modelcontextprotocol/inspector node build/index.js (or use the Genkit tester via npx genkit start -- npx tsx --watch src/index.ts).

Key features of Firebase Docs MCP Server Setup

• Indexes multiple Firebase documentation pages into a local SQLite database
• Uses Gemini embedding models for document retrieval
• Provides a stdio-based MCP server for tool-based queries
• Includes a Genkit MCP client for interactive testing
• Implements retry logic on indexing failures

Use cases of Firebase Docs MCP Server Setup

• Querying Firebase documentation through an AI assistant using MCP tools
• Building a custom documentation retrieval pipeline for Firebase content
• Testing and debugging an MCP server with a local Genkit DevUI
• Integrating Firebase docs into Genkit flows via the find-firebase-doc tool

FAQ from Firebase Docs MCP Server Setup

What API key is required?

A Gemini API key from AI Studio is required — set it with the genaikey environment variable.

Where is the indexed database stored?

The SQLite database is stored at $HOME/.indexResp/db.sqlite.

What transport does the MCP server use?

The server uses STDIO transport.

Are there any known issues with running the inspector?

The author reports trouble using the VSCode integrated terminal for the inspector; use your system terminal instead.

What dependencies are needed?

Go, Node.js, npm, and Genkit are required. The indexer uses Go and the server uses Node.js.