Autonomous Software Factory — Describe a project, watch AI agents build it in real-time.
███╗ ██╗███████╗██╗ ██╗██╗ ██╗███████╗
████╗ ██║██╔════╝╚██╗██╔╝██║ ██║██╔════╝
██╔██╗ ██║█████╗ ╚███╔╝ ██║ ██║███████╗
██║╚██╗██║██╔══╝ ██╔██╗ ██║ ██║╚════██║
██║ ╚████║███████╗██╔╝ ╚██╗╚██████╔╝███████║
╚═╝ ╚═══╝╚══════╝╚═╝ ╚═╝ ╚═════╝ ╚══════╝
You describe a software project → 5 specialized AI agents collaborate to build it:
┌──────────────┐
│ 🏗️ Architect │ ← Designs architecture
└──────┬───────┘
┌─────┴─────┐
▼ ▼
┌──────────┐ ┌──────────┐
│ ⚡ Coder │ │ 📝 Docs │ ← Code + Docs in parallel
└─────┬────┘ └──────────┘
▼
┌──────────┐
│ 🧪 Tester │ ← Generates tests
└─────┬────┘
▼
┌──────────┐
│ 🔍 Review │ ← Reviews everything
└──────────┘
All visible in a real-time TUI dashboard that shows agents thinking, coding, and collaborating.
cd nexus
uv sync
export ANTHROPIC_API_KEY=sk-ant-xxxxx
# Build a project
nexus build "A REST API for managing a todo list with SQLite persistence"
# Interactive mode
nexus interactive
# View pipeline DAG
nexus pipeline
# List agents
nexus agents
# Start API server
nexus serveThe real-time dashboard shows:
| Panel | What it shows |
|---|---|
| Agents | Status and progress of each agent |
| Pipeline | DAG execution waves |
| Event Log | Live streaming of agent activity |
| Files | Project file tree as it's built |
| Stats | Tokens, duration, completion |
nexus serve # http://localhost:8080
# Create a build
curl -X POST http://localhost:8080/builds \
-H "Content-Type: application/json" \
-d '{"name": "myapp", "description": "A chat application"}'
# Check status
curl http://localhost:8080/builds/{session_id}
# WebSocket for real-time events
wscat -c ws://localhost:8080/wssrc/nexus/
├── models.py # Core data models (Pydantic)
├── config.py # Settings (pydantic-settings)
├── event_bus.py # Async pub/sub with wildcards
├── cli.py # Click CLI with Rich output
├── agents/
│ ├── base.py # BaseNexusAgent (Plan→Execute→Reflect)
│ ├── architect.py # 🏗️ System design & file structure
│ ├── coder.py # ⚡ Code generation (parallel batches)
│ ├── tester.py # 🧪 Test suite generation
│ ├── reviewer.py # 🔍 Code review & quality assessment
│ └── docs.py # 📝 README, API docs, architecture docs
├── pipeline/
│ ├── scheduler.py # DAG topological sort → parallel waves
│ └── engine.py # Pipeline execution with events
├── tui/
│ └── dashboard.py # Rich Live real-time dashboard
├── server/
│ └── app.py # FastAPI REST + WebSocket
└── workspace/
└── manager.py # File system output manager
uv run pytest tests/ -v