Skip to main content
Glama

FireMCP šŸ”„

A Model Context Protocol (MCP) server for Firestore, enabling AI agents and LLMs to interact with Firestore databases securely through multiple transport protocols.

License: MIT

🌟 Features

  • Multiple Transport Protocols: Support for stdio, HTTP Streamable, and Server-Sent Events (SSE)

  • Secure by Design: Uses Firebase Client SDK with Firestore Security Rules instead of Admin SDK

  • Complete CRUD Operations: Get, set, add, delete, and query Firestore documents

  • Type-Safe: Built with TypeScript and Zod schemas

  • MCP Compatible: Works with any MCP-compatible client (Claude Desktop, etc.)

šŸ”’ Security Philosophy

FireMCP uses the Firebase Client SDK instead of the Admin SDK for a critical security reason:

The Client SDK respects Firestore Security Rules, ensuring that AI agents can only access data they're explicitly permitted to. This prevents unauthorized access to sensitive resources, even if the AI behaves unexpectedly.

With the Admin SDK, an AI would have unrestricted access to all Firestore data, which could be a significant security risk in production environments.

šŸ“‹ Prerequisites

  • Bun v1.0 or higher

  • A Firebase project with Firestore enabled

  • Firebase Authentication configured with at least one user

šŸš€ Quick Start

1. Clone the Repository

git clone https://github.com/iamanishroy/firemcp.git cd firemcp

2. Install Dependencies

bun install

3. Configure Environment Variables

Copy the example environment file and fill in your Firebase credentials:

cp .env.example .env

Edit .env with your Firebase configuration:

# Required FIREBASE_API_KEY=your-api-key-here FIREBASE_PROJECT_ID=your-project-id FIREBASE_USER_EMAIL=user@example.com FIREBASE_USER_PASSWORD=your-password-here # Optional PORT=3003

Note: You can find these values in your Firebase Console under Project Settings.

4. Run the Server

Choose your preferred transport protocol:

# stdio (default - for MCP clients like Claude Desktop) bun run stdio # HTTP Streamable bun run http # Server-Sent Events (SSE) bun run sse

šŸ”§ Available Tools

FireMCP provides five Firestore operations as MCP tools:

1. get_document

Retrieve a document from Firestore by its path.

Input:

{ "path": "users/userId" }

Output:

{ "exists": true, "data": { "name": "John Doe", "email": "john@example.com" } }

2. set_document

Create or overwrite a document in Firestore.

Input:

{ "path": "users/userId", "data": { "name": "Jane Doe", "email": "jane@example.com" }, "merge": false }

Output:

{ "success": true }

3. add_document

Add a new document to a collection with an auto-generated ID.

Input:

{ "collection": "users", "data": { "name": "Bob Smith", "email": "bob@example.com" } }

Output:

{ "id": "auto-generated-id" }

4. delete_document

Delete a document from Firestore.

Input:

{ "path": "users/userId" }

Output:

{ "success": true }

5. query_collection

Query a Firestore collection with filters and limits.

Input:

{ "collection": "users", "filters": [ { "field": "age", "operator": ">", "value": 18 }, { "field": "active", "operator": "==", "value": true } ], "limit": 10 }

Output:

{ "documents": [ { "id": "doc1", "name": "Alice", "age": 25, "active": true }, { "id": "doc2", "name": "Bob", "age": 30, "active": true } ] }

šŸ”Œ Integration with MCP Clients

Claude Desktop

Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{ "mcpServers": { "firestore": { "command": "bun", "args": ["run", "/path/to/firemcp/index.ts"], "env": { "FIREBASE_API_KEY": "your-api-key", "FIREBASE_PROJECT_ID": "your-project-id", "FIREBASE_USER_EMAIL": "user@example.com", "FIREBASE_USER_PASSWORD": "your-password" } } } }

MCP Inspector

For debugging and testing:

bun run inspect

šŸ› ļø Development

Scripts

  • bun run start - Start with default transport (stdio)

  • bun run dev - Same as start

  • bun run http - Start with HTTP Streamable transport

  • bun run sse - Start with SSE transport

  • bun run stdio - Start with stdio transport

  • bun run inspect - Launch MCP Inspector for debugging

  • bun run kill - Kill any process running on port 3003

Adding New Tools

  1. Create a new file in src/tools/

  2. Define input/output schemas using Zod

  3. Implement the tool function

  4. Register the tool in src/server.ts

Example:

import { z } from "zod"; import { getFirestoreInstance } from "../firestore.js"; const inputSchemaObject = z.object({ // Define your input schema }); const outputSchemaObject = z.object({ // Define your output schema }); export const inputSchema = inputSchemaObject.shape; export const outputSchema = outputSchemaObject.shape; export const myTool = async (input: z.infer<typeof inputSchemaObject>) => { const db = await getFirestoreInstance(); // Implement your tool logic return { content: [{ type: 'text' as const, text: JSON.stringify(result) }], structuredContent: result }; };

šŸ” Firestore Security Rules

Since FireMCP uses the Client SDK, you must configure Firestore Security Rules. Example rules:

rules_version = '2'; service cloud.firestore { match /databases/{database}/documents { // Allow authenticated users to read/write their own data match /users/{userId} { allow read, write: if request.auth != null && request.auth.uid == userId; } // Allow authenticated users to read public collections match /public/{document=**} { allow read: if request.auth != null; } } }

šŸ“ License

This project is licensed under the MIT License - see the LICENSE file for details.

šŸ™ Acknowledgments

šŸ“§ Support

If you have any questions or run into issues, please open an issue on GitHub.

-
security - not tested
A
license - permissive license
-
quality - not tested

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/iamanishroy/firemcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server