Web Interface / API Server Guide (v0.1.x)
Penguin currently includes a FastAPI-based HTTP server that exposes core functionality (projects, tasks, chat) as a REST/WS API. A rich graphical UI is planned (see future considerations), but not available yet.
Installation
# Install Penguin with web extras
pip install "penguin-ai[web]"
Starting the server
# Default → http://localhost:8000
penguin-web
# Custom host/port
penguin-web --host 0.0.0.0 --port 9000
The command is a thin wrapper around penguin.web.app:app (FastAPI) using uvicorn.
Optional flags:
| Flag | Description |
|---|---|
--debug | Enable reload & verbose logging |
--workers N | Run with a process pool (production) |
API Overview
The OpenAPI / Swagger UI is served at:
GET http://<host>:<port>/docs
Key route groups:
/projects– CRUD for projects/tasks– CRUD & execution for tasks/chat– Chat endpoint (POST message, returns assistant reply)/ws/chat– WebSocket for streaming chat responses
For full path details, inspect the live Swagger page or read penguin/penguin/web/routes.py.
Example – list projects:
curl http://localhost:8000/projects
Authentication
Authentication is not yet implemented. The server should only be exposed on trusted networks.
Known Limitations
- No HTML UI – interaction is via REST/WS or the CLI.
- Auth/RBAC, CORS, and HTTPS termination are future work.
- API surface may change without notice until v1.0.
Last updated: June 13th 2025