
Firebase Mcp
io.github.firebasePublished by GitHub account firebasev0.3.0Updated Oct 24, 2026
Gives AI development tools Firebase-specific capabilities and expertise.
Overview
A local MCP server that gives an AI assistant Firebase project, database, auth, messaging, Crashlytics and deployment capabilities through the Firebase CLI.
- What you can do
- Create, list and inspect Firebase projects, apps and environment settings, and sign in or out of the CLI session. Read, query, write and delete Firestore documents, and read or write Realtime Database paths. Retrieve and update Auth users, and send Cloud Messaging push notifications to tokens or topics. Validate and fetch security rules, and build or execute Firebase SQL Connect schemas and queries. Inspect Crashlytics issues, events and reports, read Cloud Functions and App Hosting logs, and manage Remote Config templates.
- When to use it
- Useful when an assistant is working inside a Firebase-backed app and should inspect or change projects, data, auth users, rules or deployment state without switching to the console or CLI by hand.
- Requirements
- Node.js and npm; runs locally via npx firebase-tools. It uses the same user credentials that authorize the Firebase CLI in the environment where it runs, so a Firebase login with access to the target projects is needed. Desktop MCP clients only.
Installation
In SourceWeft
- Open Firebase Mcp in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
Firebase MCP Server
[License] [Install MCP Server]
The Firebase Model Context Protocol (MCP) Server gives AI-powered development tools the ability to work with your Firebase projects and your app's codebase. The Firebase MCP server works with any tool that can act as an MCP client, including: Firebase Studio, Gemini CLI, Claude Code, Cline, Cursor, VS Code Copilot, Windsurf, and more!
Features
An editor configured to use the Firebase MCP server can use its AI capabilities to help you:
- Create and manage Firebase projects - Initialize new projects, list existing ones, and manage Firebase apps
- Manage Firebase Authentication users - Retrieve, update, and manage user accounts
- Work with Cloud Firestore and Firebase SQL Connect - Query, read, write, and manage database documents
- Retrieve Firebase SQL Connect schemas - Generate schemas and operations with AI assistance
- Understand security rules - Validate and retrieve security rules for Firestore, Cloud Storage, and Realtime Database
- Send messages with Firebase Cloud Messaging - Send push notifications to devices and topics
- Access Crashlytics data - Debug issues, view crash reports, and manage crash analytics
- Deploy to App Hosting - Monitor backends and retrieve logs
- Work with Realtime Database - Read and write data in real-time
- Query Cloud Functions logs - Retrieve and analyze function execution logs
- Manage Remote Config - Get and update remote configuration templates
Some tools use Gemini in Firebase to help you:
- Generate Firebase SQL Connect schema and operations
- Consult Gemini about Firebase products
Important: Gemini in Firebase can generate output that seems plausible but is factually incorrect. It may respond with inaccurate information that doesn't represent Google's views. Validate all output from Gemini before you use it and don't use untested generated code in production. Don't enter personally-identifiable information (PII) or user data into the chat.
Learn more about Gemini in Firebase and how it uses your data.
Installation and Setup
Prerequisites
Make sure you have a working installation of Node.js and npm.
Basic Configuration
The Firebase MCP server can work with any MCP client that supports standard I/O (stdio) as the transport medium. When the Firebase MCP server makes tool calls, it uses the same user credentials that authorize the Firebase CLI in the environment where it's running.
Here are configuration instructions for popular AI-assistive tools:
Gemini CLI
Install the Firebase extension for Gemini CLI:
Claude Code
Option 1: Install via plugin (Recommended)
The easiest way to set up the Firebase MCP server in Claude Code is to install the official Firebase plugin:
-
Add the Firebase marketplace for Claude plugins:
-
Install the Claude plugin for Firebase:
-
Verify the installation:
Option 2: Configure MCP server manually
Alternatively, you can manually configure the Firebase MCP server by running:
You can verify the installation by running:
It should show:
Cursor
Add to .cursorrules in your project directory or configure in Cursor settings:
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
Firebase Studio
To configure Firebase Studio to use the Firebase MCP server, edit or create the configuration file: .idx/mcp.json
Usage
Once configured, the MCP server will automatically provide Firebase capabilities to your AI assistant. You can:
- Ask the AI to help set up Firebase services
- Query your Firestore database
- Manage authentication users
- Deploy to Firebase Hosting
- Debug Crashlytics issues
- And much more!
For a complete list of available tools and resources, see the Server Capabilities section below.
Documentation
For more information, visit the official Firebase MCP server documentation.
Server Capabilities
The Firebase MCP server provides three types of capabilities: Tools (functions that perform actions), Prompts (reusable command templates), and Resources (documentation files for AI models).
Source: src/mcp/README.md at commit c9c2461
Tools
0Version history
1- v0.3.0LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts