bgagents
Enables turning AI code agents like Anthropic Claude and OpenAI Codex into background agents accessible via MCP protocol for code generation, branch creation, and PR automation.
README
bgagents

bgagents is the first open-source project that turns the most powerful code agents, such as ANTHROPIC ClaudeCode and OpenAI Codex into background agents, accessible via REST API, MCP, and an interactive frontend
⨠Features
Getting Started
Quick Start (Recommended)
The fastest way to get bgagents running is with our pre-built Docker image:
GitHub Authentication (Choose one method):
Option 1: Direct Token
docker run -e GITHUB_TOKEN=[ghp_xxx] \
-e API_KEY=[sk-...-xxx] \ # Optional
bgagents/bgagents:v1.0.1-alpha
Option 2: OAuth Flow
docker run -e GITHUB_CLIENT_ID=[your_client_id] \
-e GITHUB_CLIENT_SECRET=[your_client_secret] \
-e GITHUB_REDIRECT_URI=http://localhost:3000 \
-e API_KEY=[sk-...-xxx] \ # Optional
bgagents/bgagents:v1.0.1-alpha
Optional Environment Variables:
API_KEY: Your Anthropic Claude or OpenAI API key (auto-detects provider)GITHUB_TOKEN: GitHub access token (alternative to providing it in requests)
š¤ Supported AI Models:
- Anthropic: claude-sonnet-4-20250514 (default), claude-opus-4-20250514, claude-3-7-sonnet-20250219
- OpenAI: codex-mini-latest (default), o4-mini
Production
Prerequisites
With Docker Compose (Recommended)
-
Clone the repository:
git clone https://github.com/michael-elkabetz/claudecodex.git cd claudecodex -
Set up environment variables:
Update
docker-compose.ymlwith the following environment variables:VITE_GITHUB_CLIENT_ID=[VITE_GITHUB_CLIENT_ID] VITE_GITHUB_REDIRECT_URI=[VITE_GITHUB_REDIRECT_URI] GITHUB_CLIENT_SECRET=[GITHUB_CLIENT_SECRET] GITHUB_TOKEN=ghp_xxx API_KEY=[your_anthropic_or_openai_api_key] # Optional -
Deploy with Docker Compose:
# Build and start all services docker-compose up --build -d # View logs docker-compose logs -f -
Access the services:
- Frontend: http://localhost (port 80)
- Backend API: http://localhost:3000
- MCP Server: http://localhost:6213
With Docker (Frontend + Backend)
-
Clone the repository:
git clone https://github.com/michael-elkabetz/claudecodex.git cd claudecodex -
Set up environment variables:
Update the main
Dockerfilewith the following environment variables:VITE_GITHUB_CLIENT_ID=[VITE_GITHUB_CLIENT_ID] VITE_GITHUB_REDIRECT_URI=[VITE_GITHUB_REDIRECT_URI] GITHUB_CLIENT_SECRET=[GITHUB_CLIENT_SECRET] GITHUB_TOKEN=ghp_xxx API_KEY=[your_anthropic_or_openai_api_key] # Optional
3. Build and deploy:
```bash
docker build -t ClaudeCodex-backend .
docker run -d -p 80:80 ClaudeCodex-backend
```
Development
Quick Setup
npm install
npm run dev
# Individual services
npm run dev:frontend
npm run dev:backend
npm run dev:mcp
Development URLs
- Frontend: http://localhost
- Backend: http://localhost:3000
- MCP Server: http://localhost:6213
š Docker Deployment
Using Docker Compose (Recommended)
# Production deployment
docker-compose up --build -d
# Development with hot reload
docker-compose -f docker-compose.dev.yml up --build
Individual Services
# Frontend
docker build -f frontend/Dockerfile -t claudecodex-frontend ./frontend
docker run -d -p 80:80 claudecodex-frontend
# Backend
docker build -f backend/Dockerfile -t ClaudeCodex-backend ./backend
docker run -d -p 3000:3000 ClaudeCodex-backend
# MCP Server
docker build -f mcp/Dockerfile -t ClaudeCodex-mcp ./mcp
docker run -d -p 6213:6213 ClaudeCodex-mcp
Docker Compose Services
| Service | Port | Description |
|---|---|---|
| frontend | 80 | React TypeScript UI |
| backend | 3000 | Express.js REST API |
| mcp | 6213 | Model Context Protocol Server |
š MCP Integration
Connecting to MCP Server
The MCP server runs on port 6213 and provides AI-powered code generation capabilities.
Connection URL: http://localhost:6213
MCP Inspector Setup
-
Start the MCP Inspector:
npm run inspector -
Open Inspector UI:
- Navigate to: http://127.0.0.1:6274
-
Configure Connection:
- Transport Type:
Streamable HTTP - MCP Server URL:
http://localhost:6213
- Transport Type:
-
Authentication:
- Copy the
MCP_PROXY_AUTH_TOKENfrom the inspector logs - Paste it in the Proxy Session Token field
- Click Connect
- Copy the
MCP Tools Available
- Execute: Complete AI-powered workflow (code generation, branch creation, PR automation)
- Integrates with GitHub repositories
- Supports Anthropic Claude and OpenAI GPT
- Automated branch and PR creation
š API Documentation
Backend Endpoints
- Swagger UI: http://localhost:3000/api-docs
- Health Check:
GET /api/health - AI Code Generation:
POST /api/dev/execute - GitHub Authentication:
POST /api/github/auth - Repository Branches:
POST /api/github/branches
MCP Protocol
- Server Info:
GET / - Tools List:
GET /tools - Generate:
POST /tools/generate
š ļø Development
Tech Stack
Project Structure
ClaudeCodex/
āāā frontend/ # React TypeScript UI
āāā backend/ # Express.js API
āāā mcp/ # MCP Server
āāā docker-compose.yml # Multi-service deployment
āāā README.md # You are here
Available Scripts
npm run dev # Start all services
npm run build # Build all services
npm run test # Run all tests
npm run lint # Lint all projects
npm run clean # Clean build artifacts
npm run dev:frontend # Frontend development
npm run dev:backend # Backend development
npm run dev:mcp # MCP server development
š¤ Contributing
- Fork the repository
- Create your feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
š License
This project is licensed under the MIT License - see the LICENSE file for details.
š Acknowledgments
- Thanks to Anthropic for Claude Code
- Thanks to OpenAI for Codex
- Thanks to the open-source community
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.