An AI-powered coding assistant that can control your computer, write code, debug issues, and automate development tasks.
- AI Code Generation - Generate code in any language with natural language descriptions
- Intelligent Debugging - Analyze and fix bugs with AI assistance
- Code Refactoring - Improve code quality and performance automatically
- Test Generation - Create comprehensive test suites
- Documentation - Generate API docs, READMEs, and code comments
- Computer Control - AI can control desktop, run commands, and use IDEs
- Full Monitoring - Production-ready logging, metrics, and alerting
- Docker & Docker Compose
- Anthropic API Key (Get one here)
- 8GB+ RAM recommended
- 20GB+ free disk space
Edit .env
file and add your Anthropic API key:
ANTHROPIC_API_KEY=sk-ant-api03-xxxxx...
./scripts/start.sh
- Main UI: http://localhost:8080
- API Docs: http://localhost:8000/docs
- VNC Desktop: http://localhost:6080/vnc.html
- Code Server: http://localhost:8443
- Grafana: http://localhost:3000 (admin/admin)
- Open http://localhost:8080
- Enter your coding task
- Select language and preferences
- Click "Generate Code"
- View results and download code
import requests
response = requests.post(
"http://localhost:8000/execute",
json={
"task": "Create a Python REST API for todo management",
"context": {"language": "python", "framework": "fastapi"}
}
)
print(response.json())
curl -X POST "http://localhost:8000/execute" \
-H "Content-Type: application/json" \
-d '{"task": "Write a sorting algorithm in Python"}'
See .env
file for all configuration options:
ANTHROPIC_API_KEY
- Your API key (required)LOG_LEVEL
- Logging level (INFO, DEBUG, ERROR)VNC_PASSWORD
- VNC access password- Alert configurations (Email, Slack, PagerDuty)
Service | Port | Description |
---|---|---|
Streamlit | 8080 | Main web interface |
FastAPI | 8000 | REST API |
VNC | 5900 | VNC desktop access |
noVNC | 6080 | Web-based VNC |
Code Server | 8443 | VS Code in browser |
Grafana | 3000 | Monitoring dashboards |
Prometheus | 9091 | Metrics |
- Grafana: http://localhost:3000
- Username: admin
- Password: admin (change on first login)
- Prometheus: http://localhost:9091
- Jaeger: http://localhost:16686 (tracing)
- Request rate and latency
- Error rates by type
- Task completion times
- System resources (CPU, memory, disk)
- Active sessions
Configure alerts in .env
:
- Email alerts via SMTP
- Slack webhooks
- PagerDuty integration
./scripts/start.sh
./scripts/stop.sh
docker-compose logs -f
./scripts/update.sh
./scripts/backup.sh
docker exec -it computer-use-coding bash
computer-use-coding-assistant/
ββ
βββ agent/ # AI agent logic
βββ api/ # FastAPI server
βββ monitoring/ # Monitoring tools
βββ docker/ # Docker configuration
βββ monitoring/ # Monitoring configs
β βββ grafana/ # Grafana dashboards
β βββ prometheus/ # Prometheus config
βββ projects/ # Project files
βββ logs/ # Application logs
βββ scripts/ # Utility scripts
βββ docker-compose.yml # Service orchestration
βββ .env # Configuration
βββ README.md # Documentation
- Verify key in
.env
file - Check key has sufficient credits
- Ensure key is active
- Modify ports in
docker-compose.yml
- Check for conflicting services
# Check container status
docker-compose ps
# View logs
docker-compose logs coding-assistant
# Restart services
docker-compose restart
- Allocate more Docker resources
- Check system requirements
- Monitor via Grafana dashboards
- Run in isolated Docker environment
- VNC password protected
- API key stored securely
- Automatic log redaction
- Network isolation
- Check logs:
docker-compose logs -f
- API documentation: http://localhost:8000/docs
- Monitoring dashboard: http://localhost:3000
MIT License - See LICENSE file for details
Built with: