Firebase MCP: Bridge Your AI Assistant to Firebase in Minutes

Stop writing custom API endpoints every time you need your AI assistant to interact with Firebase. This MCP server gives Claude, Cursor, and VS Code direct access to your Firestore databases, Storage buckets, and Authentication users through natural language.

The Firebase Integration You've Been Waiting For

Most developers building AI-powered apps hit the same wall: your AI assistant can help you write code, but it can't actually interact with your Firebase backend. You end up writing one-off scripts or API endpoints just to let your AI read a document, upload a file, or check user data.

Firebase MCP eliminates that friction entirely. Install it once, and your AI assistant can:

• Query and modify Firestore collections through conversation
• Upload files to Storage from URLs or raw content
• Look up user authentication data
• Handle complex database operations without you writing a single API endpoint

Real-World Use Cases That Save Hours

Content Management Workflows "Upload this image to Firebase Storage and add a document to the 'posts' collection with the download URL" - Your AI handles the entire pipeline, no custom upload scripts needed.

Customer Support Integration
"Find all users who signed up in the last week and check their profile completeness" - Direct database queries through natural language, perfect for support dashboards.

Data Migration and Cleanup "List all documents in the 'old_users' collection and move active ones to 'users' with updated schema" - Complex batch operations handled conversationally.

Development and Testing "Create test data for the product catalog with 20 sample items" - Populate development databases without writing seed scripts.

Installation Is Genuinely Simple

Add this to your MCP client config and you're running:

{
  "firebase-mcp": {
    "command": "npx",
    "args": ["-y", "@gannonh/firebase-mcp"],
    "env": {
      "SERVICE_ACCOUNT_KEY_PATH": "/path/to/serviceAccountKey.json",
      "FIREBASE_STORAGE_BUCKET": "your-project.firebasestorage.app"
    }
  }
}

That's it. Your AI assistant now has admin access to your Firebase project through 11 different tools covering Firestore operations, Storage management, and user authentication.

Built for Real Development Workflows

The server includes HTTP transport support for multi-client scenarios, comprehensive error handling for common Firebase issues (like missing composite indexes), and detailed logging for debugging. It works with Firebase emulators for testing and includes proper session management.

The tool selection is practical: document CRUD operations, collection group queries, file uploads from both content and URLs, and user lookups by ID or email. These cover the majority of Firebase operations you'd actually want an AI to handle.

Works With Your Existing Setup

Compatible with Claude Desktop, VS Code Copilot, Cursor, and Augment Code out of the box. Supports both stdio and HTTP transport modes depending on your architecture. The server handles Firebase Admin SDK initialization and provides proper error messages when things like composite indexes are missing.

If you're building AI features on Firebase, this server eliminates the tedious integration work and lets you focus on the actual AI functionality. Your assistant becomes a natural interface to your Firebase backend instead of just a code writing tool.

The 172 GitHub stars and active development suggest this isn't just solving problems for early adopters - it's becoming essential infrastructure for AI-Firebase integration.