Connect Firebase to AI: Let Claude and Cursor Access Firestore, Auth, and Cloud Functions

Why Connect Firebase to AI?

Firebase powers your app's backend. Firestore documents, Authentication users, Cloud Functions logs, Hosting deployments. When you're debugging a user issue or checking data in the terminal, you open the Firebase Console, navigate through nested collections, and manually look up the values you need.

With an MCP server pointed at your Firebase Console, AI agents can query your project directly. Ask about a user's auth state, check a Firestore document, or pull recent function invocations without switching windows.

What AI Agents Can Do With Firebase Access

Once connected, Claude, Cursor, or Codex can:

• Query Firestore — "Show me the document for user abc123 in the orders collection"
• Check auth users — "What providers are linked to this email address?"
• Monitor functions — "What errors occurred in the processPayment function today?"
• Inspect rules — "What are the current Firestore security rules?"
• Review hosting — "When was the last deployment and what domain is it on?"
• Check usage — "How many Firestore reads have we used this billing period?"

How It Works

1. Open DataFaucet and enter your Firebase Console URL
2. Browse your project — click through Firestore collections, Auth users, Functions logs
3. DataFaucet captures the read operations as MCP tools
4. Deploy — your AI agents now have structured access to query Firebase

You're browsing the Firebase Console web UI, not sharing service account credentials. DataFaucet captures the HTTP interactions and replays them as tool calls. Your service accounts and admin SDK keys stay where they are.

Example Queries

"What does the user profile document look like for uid abc123?"
"How many users signed up in the last 24 hours?"
"Show me the last 5 function invocations that returned errors"
"What collections exist under the /organizations path?"
"Is the user with email test@example.com verified?"

When This Saves Time

• User support — look up user state without opening Console during a call
• Debugging — check Firestore document structure while writing queries in code
• Incident response — pull function error logs alongside your code changes
• Data verification — confirm what's actually in Firestore vs what your code expects

Security Model

DataFaucet connects through the Firebase Console web interface. Access is scoped to whatever your Google account can see in the Console. Read-only operations mean agents query but don't modify data. You control exposure by which Console pages you browse during setup.

Getting Started

Point DataFaucet at your Firebase Console URL (console.firebase.google.com/project/your-project). Browse through Firestore, Auth, and Functions. Deploy. Your agents can now query Firebase project data as part of their normal workflow.

Create your Firebase MCP server →