MCP 服务器

Brainbase MCP Server

Provides 70 tools to interact with the Brainbase API, enabling management of workers, chat/voice deployments, flows, resources, and more via natural language.

README

Brainbase MCP Server

A Model Context Protocol (MCP) server that provides comprehensive access to the Brainbase API.

Features

This MCP server exposes 70 tools covering all aspects of the Brainbase API:

🤖 Workers (5 tools)

Create, read, update, and delete workers
List all workers for your team

💬 Chat Deployments (6 tools)

Manage chat deployments
Deploy and configure chat agents
Query by deployment ID or agent ID

🎤 Voice Deployments (11 tools)

Create and manage voice deployments
Configure custom webhooks
Make batch voice calls
Get deployment analytics

🎤 Voice V1 Deployments (11 tools)

Legacy voice deployment management
Campaign management
Batch calling capabilities

📊 Flows (5 tools)

Create and manage conversation flows
Update flow definitions
List flows by worker

📁 Folders (6 tools)

Organize resources with folders
Hierarchical folder structure
Move resources between folders

📄 Resources (4 tools)

Manage file and link resources
Vector search (RAG) capabilities
Move and organize resources

📎 File Resources (2 tools)

Upload and manage file resources
Associate files with workers

🔗 Link Resources (2 tools)

Create and manage link resources
Web content integration

🔌 Integrations (4 tools)

Twilio integration management
Configure external services

👥 Team (1 tool)

Get team information

🧪 Tests (5 tools)

Create and manage tests
Run test suites
View test run history

📞 Assets (3 tools)

Manage phone numbers
Register Twilio numbers

📊 Logs & Analysis (5 tools)

Chat deployment logs
Voice deployment logs
Voice analytics with billing breakdown

Installation

Using pip

pip install git+https://github.com/brainbase-mcp/brainbase-mcp.git

Using uv (recommended)

uv pip install git+https://github.com/brainbase-mcp/brainbase-mcp.git

From source

git clone https://github.com/brainbase-mcp/brainbase-mcp.git
cd brainbase-mcp
pip install -e .

Configuration

Get your Brainbase API Key

Go to Brainbase
Sign in to your account
Navigate to Settings → API Keys
Create a new API key or copy your existing key

Configure Claude Desktop

Add this to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "brainbase": {
      "command": "python",
      "args": ["-m", "server"],
      "env": {
        "BRAINBASE_API_KEY": "your-api-key-here"
      }
    }
  }
}

Alternative: Using uv

{
  "mcpServers": {
    "brainbase": {
      "command": "uv",
      "args": ["--directory", "/path/to/brainbase-mcp", "run", "server.py"],
      "env": {
        "BRAINBASE_API_KEY": "your-api-key-here"
      }
    }
  }
}

Usage

Once configured, the Brainbase tools will be available in Claude Desktop. You can:

Managing Workers

Create a new worker named "Customer Support Bot" with description "Handles customer inquiries"

List all workers for my team

Get details for worker with ID "worker_123"

Creating Chat Deployments

Create a chat deployment for worker "worker_123" named "Website Chat" using flow "flow_456"

List all chat deployments for worker "worker_123"

Voice Deployments

Create a voice deployment for worker "worker_123" named "Sales Calls" using flow "flow_789"

Make batch calls with voice deployment "deploy_123" to these numbers: +1234567890, +1987654321

Managing Resources

Create a file resource named "Product Manual" for worker "worker_123" with URL "https://example.com/manual.pdf"

Query resources for worker "worker_123" with query "pricing information"

Working with Flows

Create a new flow for worker "worker_123" named "Support Flow" with description "Customer support conversation flow"

Update flow "flow_123" with new definition

Organizing with Folders

Create a folder named "Documentation" for worker "worker_123"

Move resource "resource_123" to folder "folder_456"

API Coverage

This MCP server implements all 70 endpoints from the Brainbase API v2:

Category	Endpoints
Workers	5
Chat Deployments	6
Voice Deployments	11
Voice V1 Deployments	11
Flows	5
Folders	6
Resources	4
File Resources	2
Link Resources	2
Integrations	4
Team	1
Tests	5
Assets	3
Logs	4
Analysis	1
Total	70

Development

Running Tests

# Install dev dependencies
pip install pytest pytest-asyncio

# Run tests
pytest test_server.py -v

Project Structure

brainbase-mcp/
├── server.py           # Main MCP server implementation
├── test_server.py      # Comprehensive test suite
├── pyproject.toml      # Project configuration
└── README.md           # This file

API Documentation

For detailed information about each endpoint, parameters, and response formats, see the official Brainbase API Documentation.

Authentication

All requests are authenticated using your Brainbase API key via the x-api-key header. The API key must be set in the BRAINBASE_API_KEY environment variable.