ClawRun is a hosting and lifecycle layer for open-source AI agents. It deploys agents into secure sandboxes (Vercel Sandbox, with more providers coming) and manages their full lifecycle, including startup, heartbeat keep-alive, snapshot/resume, and wake-on-message.

Learn more at clawrun.sh.

• 🚀 Deploy any supported AI agent with a single command
• 💤 Persistent sandboxes that sleep when idle and wake on message
• 💬 Connect messaging channels like Telegram, Discord, Slack, WhatsApp, and more
• 🖥️ Web dashboard and CLI for real-time chat and management
• 💰 Cost tracking and budget enforcement across all channels
• 🔌 Pluggable architecture for agents, providers, and channels

The deploy wizard walks you through:

1. Choosing an LLM provider and model
2. Configuring messaging channels
3. Setting cost limits and network policy
4. Deploying to your chosen provider (e.g. Vercel)

Once deployed, chat with your agent from the terminal:

clawrun agent my-instance

Or open the web dashboard:

For full setup guides, framework examples, and configuration reference, see the docs.

Report issues and suggest improvements on GitHub: Issues.

Join the community: Discussions.

Apache-2.0