smolagentsDeploy smolagents to production
Crewship is adding support for Hugging Face smolagents. Deploy code-first AI agents that use 30% fewer LLM calls. Get production API endpoints with secure sandboxed execution.
One-Command Deploy
Deploy with `crewship deploy`. No Docker or Kubernetes required.
Code-First Agents
Native support for CodeAgent that writes and executes Python.
Auto-scaling
Scale to zero when idle, scale up automatically under load.
Secure Execution
Sandboxed code execution with E2B, Docker, or WebAssembly.
What we're building
smolagents support will bring lightweight, code-first agents to Crewship. Deploy agents that write and execute Python code with minimal LLM overhead.
- Full smolagents framework compatibility
- CodeAgent and ToolCallingAgent support
- Secure sandboxed code execution
- Multi-modal support (text, vision, audio)
- Hub integration for sharing agents
$ crewship deploy
Detecting framework... smolagents
Building container...
Deploying to production...
✓ Deployed to my-agent
✓ API endpoint ready
✓ Sandboxed execution enabled
https://api.crewship.dev/v1/runsWhat is smolagents?
smolagents is Hugging Face's lightweight library for building AI agents. With 25,000+ GitHub stars, it's designed for simplicity—the core logic is just ~1,000 lines of code.
Unlike traditional tool-calling approaches, smolagents uses a "code-first" approach where agents write Python code to perform tasks. This results in 30% fewer LLM calls compared to JSON-based tool calling.
Crewship will handle the deployment and secure execution of your smolagents, so you can focus on building efficient agents while we handle the infrastructure.
Get notified when smolagents support launches
Join the waitlist and be the first to deploy smolagents on Crewship.