Skip to main content

Develop with AI

Enhance your AI-powered development workflow with Firebase Hosting. Use our context files and MCP server to give your AI tools deep Firebase knowledge and direct access to Firebase CLI functionality.

On this page​

  • Context Files - Download instruction files to make your AI smarter about Firebase Hosting
  • MCP Server - Connect AI tools directly to Firebase CLI for automated deployments
  • Additional Resources - Access Firebase LLM docs and configuration guides

Firebase Hosting Context Files​

Download our specialized instruction file to help your AI assistant generate accurate Firebase Hosting code:

Loading AI instructions...

What You Get​

  • Modern Firebase Hosting patterns and best practices
  • Accurate configuration examples
  • Performance optimization techniques
  • Common troubleshooting solutions

How to Use​

  1. Download the file above for your AI assistant
  2. Add it to your project root (e.g., gemini.md or claude.md)
  3. Reference it when asking about Firebase Hosting

Firebase CLI MCP Server​

Connect your AI directly to Firebase CLI functionality. The MCP (Model Context Protocol) server lets AI assistants execute Firebase commands and manage your projects.

What is MCP?​

MCP enables AI tools to interact with external systems through a standardized protocol. With Firebase's MCP server, your AI can:

  • Deploy sites directly to Firebase Hosting
  • Manage configurations and update firebase.json
  • List deployments and rollback versions
  • Create preview channels for testing
  • Query project information and status

Setup with Gemini CLI​

Configure Gemini CLI to use Firebase MCP server:

// .gemini/config.json
{
  "mcpServers": {
    "firebase": {
      "command": "npx",
      "args": ["-y", "firebase-tools@latest", "experimental:mcp"]
    }
  }
}

Now in Gemini CLI you can:

> Deploy my site to Firebase Hosting
> Show me recent deployments
> Create a preview channel for testing
> Update my caching rules for better performance

Setup with Other AI Tools​

The same MCP server works with:

  • Claude Desktop - Add to claude_desktop_config.json
  • Cursor - Configure in settings
  • VS Code Copilot - Add to workspace settings
  • Windsurf - Include in project config

Full MCP Documentation →

Additional Resources​

Firebase LLM Documentation​

Access all Firebase docs in AI-friendly format: firebase.google.com/docs/llms.txt

Learn More​

Pro Tip

Combine context files with MCP server for the best experience. Context files provide knowledge, while MCP provides direct Firebase access.