FAQ
What is Goalrail?
A framework that runs AI agents behind one interface. It wraps Claude Code, Codex, and Pi, plus custom agents you write in YAML, and gives each one a server, a UI, sandboxing, and policies.
Which models can I use?
Bring your own: an API key, a Claude or ChatGPT plan, an OpenAI- or Anthropic-compatible gateway (OpenRouter, LiteLLM, Ollama, Azure, vLLM), or a local model endpoint.
How do I run my own agent?
Write a short YAML file with a prompt and a harness, then run goalrail run on it. See the custom agent guide.
Is it safe to let an agent run on my machine?
Every command runs in an OS-level sandbox (bwrap on Linux, seatbelt on macOS), and policies can pause, block, or cap what an agent does.
Do I need a cloud provider?
No. Goalrail is open source and runs on your own machine and models. A compatible API gateway is optional, not required.
Is it ready for production?
No. It is alpha. Expect rough edges, and tell us where it breaks on Telegram.