接入你的智能体
从发布者处安装。不是 PlugMCP 端点。
远程 URL 还是本机命令? 两种安装方式的区别
All-in-One Docker deployment for Fabric - AI augmentation framework with Web UI, REST API, and MCP Server
# Clone the repository
git clone https://github.com/neosun100/fabric-docker.git
cd fabric-docker
# Copy environment file
cp .env.example .env
# Start with Docker Compose
docker compose up -d
# Access the services
# Web UI: http://localhost:5173
# REST API: http://localhost:8180
# Swagger: http://localhost:8180/swagger/index.html
# MCP: http://localhost:8181
Clone the repository
git clone https://github.com/neosun100/fabric-docker.git
cd fabric-docker
Configure environment
cp .env.example .env
# Edit .env and add your API keys
Start services
docker compose up -d
Verify deployment
# Check container health
docker ps
# Test API
curl http://localhost:8180/patterns/names
# Pull or build the image
docker build -t fabric:latest .
# Run the container
docker run -d \
--name fabric \
-p 5173:8080 \
-p 8180:8180 \
-p 8181:8181 \
-e OPENAI_API_KEY=your-key-here \
-e ANTHROPIC_API_KEY=your-key-here \
-v fabric-config:/root/.config/fabric \
fabric:latest
services:
fabric:
build:
context: .
dockerfile: Dockerfile
image: neosun/fabric:latest
container_name: fabric
restart: unless-stopped
ports:
- "5173:8080" # Web UI
- "8180:8180" # REST API
- "8181:8181" # MCP Server
environment:
- OPENAI_API_KEY=${OPENAI_API_KEY:-}
- ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY:-}
- DEFAULT_MODEL=gpt-4o
volumes:
- fabric-config:/root/.config/fabric
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8180/patterns/names"]
interval: 30s
timeout: 10s
retries: 3
volumes:
fabric-config:
Access the web interface at http://localhost:5173:
# List all patterns
curl http://localhost:8180/patterns/names
# Get a specific pattern
curl http://localhost:8180/patterns/summarize
# Check pattern exists
curl http://localhost:8180/patterns/exists/summarize
# Get available models
curl http://localhost:8180/models/names
# Get configuration
curl http://localhost:8180/config
# Get strategies
curl http://localhost:8180/strategies
The MCP server enables AI assistants (like Claude Desktop) to use Fabric patterns directly.
Configure in your MCP client:
{
"mcpServers": {
"fabric": {
"url": "http://localhost:8181"
}
}
}
┌─────────────────────────────────────────────────────────┐
│ Docker Container │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │
│ │ Web UI │ │ REST API │ │ MCP Server │ │
│ │ (Node) │ │ (Go) │ │ (Python) │ │
│ │ :8080 │ │ :8180 │ │ :8181 │ │
│ └─────────────┘ └─────────────┘ └─────────────┘ │
│ │ │ │ │
│ └────────────────┼────────────────┘ │
│ │ │
│ ┌───────────┴───────────┐ │
│ │ Fabric Core │ │
│ │ (233+ Patterns) │ │
│ └───────────────────────┘ │
└─────────────────────────────────────────────────────────┘
│ │ │
Port 5173 Port 8180 Port 8181
fabric-docker/
├── Dockerfile # Multi-stage build
├── docker-compose.yml # Compose configuration
├── docker/
│ ├── entrypoint.sh # Container entrypoint
│ ├── supervisord.conf # Process manager config
│ ├── mcp_server.py # MCP server implementation
│ └── nginx.conf # Nginx config (optional)
├── data/
│ ├── patterns/ # 233+ AI patterns
│ └── strategies/ # Prompt strategies
├── web/ # Svelte Web UI
├── cmd/ # Go commands
├── internal/ # Go internal packages
└── docs/ # Documentation
# Check logs
docker logs fabric
# Restart fabric-api service
docker exec fabric supervisorctl restart fabric-api
# Verify ports
docker exec fabric netstat -tlnp
# Check if services are running
docker exec fabric supervisorctl status
# Manual restart
docker compose restart
Edit .env to change port mappings:
PORT=8280
WEB_PORT=5273
MCP_PORT=8281
Contributions are welcome! Please feel free to submit a Pull Request.
git checkout -b feature/AmazingFeature)git commit -m 'Add some AmazingFeature')git push origin feature/AmazingFeature)This project is licensed under the MIT License - see the LICENSE file for details.

All-in-One Docker deployment for Fabric - AI augmentation framework with Web UI, REST API, and MCP Server
工具
暂无工具列表。远程服务器在首次健康检查后会获得工具列表。
常见问题
在本页复制适用于 Cursor、Claude Desktop、Claude Code 或 VS Code 的 stdio 配置。它使用发布者的安装方式,而不是 PlugMCP 代理。
此记录未列出必需的环境变量。
Fabric Docker 通过 stdio 在本地运行。
上次检查时间:2026-10-28。 All-in-One Docker deployment for Fabric - AI augmentation framework with Web UI, REST API, and MCP Server