Firebase Docs MCP

🚀 Firebase Documentation MCP Server Setup Guide

This guide provides a detailed introduction to the Firebase Documentation MCP server, including its directory layout and usage methods.

📁 Directory Layout

docs-mcp

This directory corresponds to the indexer for Firebase documentation. It's a Go project that fetches and indexes Firebase documentation from specified file paths.

docs-mcp-server

This is the Model Context Protocol server, which provides content services through standard input-output transmission.

genkit-mcp-tester

This is an MCP client implemented using GenKit, used to test the docs-mcp-server.

💻 Usage Examples

🚀 Start Indexing

1. Set the API key: We use the Gemini embedding model to process documents, so you need to obtain an API key from AI Studio. To set the key, call export genaikey="APIKEY" in the terminal.
2. Ensure the output directory is empty: We'll write files to a folder named .indexResp in the user's home directory. Go will fetch documents from the Firebase documentation website, save them in Markdown format to disk, and index them in an SQL Lite database. If indexing fails, it will implement a retry strategy to re - index the documents in Markdown format.
3. Call the indexing process: In the docs-mcp folder, call go run .. This will start the indexing process for the file paths listed near line 291 of the main.go file.

🧪 Test the Indexer

1. Set the API key: As before, obtain an API key from AI Studio and set it by calling export genaikey="APIKEY" in the terminal.
2. Switch to the docs-mcp-server folder.
3. Copy the index database: Copy the index database to the current docs-mcp-server folder by calling cp $HOME/.indexResp/db.sqlite ..
4. Install dependencies and build the project: Run npm ci and then npm run build. After the build is complete, test the project by running npm run build && npx @modelcontextprotocol/inspector node build/index.js. This will start the inspector and print a URL for you to view the STDIO server.
5. Connect and test: In the inspector view, click "Connect", then select "List Tools -> find-firebase-doc" from the tool menu. Enter the request content you want to use for testing.

⚠️ Important Note

The author found that similar issues may occur when using the VSCode terminal. If you encounter similar problems, try using the system terminal.

🧪 Test with Genkit

1. Set the API key in the code: Modify the following content in the docs-mcp-server/src/helpers/embeddings.ts file:

   const genAiKey = process.env.genaikey || "";
   

   to

   const genAiKey = process.env.genaikey || "MYAPIKEY";
   

2. Switch to the genkit-mcp-tester directory.
3. Copy the index database: Copy the index database to the current genkit-mcp-tester folder by calling cp $HOME/.indexResp/db.sqlite ..
4. Install dependencies and build the project: Run npm ci and then npm run build. After the build, start the Genkit test by running the command:

   npm start
   

5. Start testing in the terminal: Enter the following in the terminal to start testing:

   mcp-tester
   

6. Input test content: Enter the content to be processed according to the prompts for testing.