Quickstart
Get your first Clawpy swarm running in under 5 minutes. This guide covers the self-hosted Docker deployment path.
Prerequisites
Before you begin, make sure you have:
- Docker (version 20.10 or later) with Docker Compose installed and running
- At least one LLM API key (OpenAI, Anthropic, Google, or DeepSeek)
That's it. No Python installation needed — everything runs inside the container.
Step 1: Clone the Repository
git clone https://github.com/daemon168/Clawpy-dev.git
cd Clawpy-dev
Step 2: Configure Your API Keys
Create a .env file in the project root and add your preferred LLM provider key:
# Required — at least one LLM provider
OPENAI_API_KEY=sk-...
# ANTHROPIC_API_KEY=sk-ant-...
# GOOGLE_API_KEY=AIza...
# DEEPSEEK_API_KEY=sk-...
Clawpy uses a Bring Your Own Key (BYOK) model — you only pay your LLM provider directly. There are no Clawpy platform fees for self-hosted deployments.
Step 3: Launch the Platform
Start Clawpy with Docker Compose:
docker compose up -d
This spins up two services:
- Clawpy — the dashboard, API, and agent runtime on port 7420
- Ollama — local model serving for offline/private workloads
Once the containers are healthy, open your browser:
http://localhost:7420
Step 4: Complete the Onboarding Interview
On first boot, Clawpy runs a guided interview to configure your agent's identity:
- Identity — choose a role preset (DevOps Lead, Research Assistant, Code Architect, Project Manager, or Custom)
- Guardrails — select a safety profile (Strict, Balanced, Permissive, or Custom hard limits)
- Heartbeat Tasks — configure proactive background tasks that run on a schedule
Your answers are saved to soul.md and heartbeat.md — both are plain text files you can edit at any time.
Step 5: Send Your First Task
In the operator chat panel, try a simple task:
Research the top 3 Python web frameworks and create a comparison table
The swarm conductor will classify the task, assign it to the appropriate agent archetype, and execute it. You can watch the agents work in real time from the dashboard hierarchy view.
Step 6: Configure Authentication (Optional)
For production or remote deployments, secure your dashboard with a token:
# Add to your .env file
CLAWPY_DASHBOARD_TOKEN=your-secret-token
For public-facing deployments, use the docker-compose.public.yml override which adds a Caddy reverse proxy with automatic HTTPS:
docker compose -f docker-compose.yml -f docker-compose.public.yml up -d
Environment Variables
| Variable | Default | Purpose |
|---|---|---|
CLAWPY_PORT | 7420 | Dashboard and API port |
CLAWPY_ENV | production | Runtime environment |
CLAWPY_DASHBOARD_TOKEN | (none) | Optional dashboard access token |
CLAWPY_VOICE_RUNTIME | pipecat | Voice interface runtime (pipecat or legacy) |
CLAWPY_OLLAMA_BASE_URL | http://ollama:11434 | Ollama endpoint for local models |
TZ | UTC | Container timezone |
Next Steps
- Chat — Learn how the operator interface works
- Agents — Understand the different agent archetypes
- Memory — Explore how Clawpy remembers and learns
- Scheduled Runs — Set up autonomous background tasks
- Why Clawpy — See how Clawpy compares to other agent frameworks