Astro AI

About Astro AI

Astro AI is agent-native infrastructure for production AI. Define agents as code, deploy with a single command, and get models, knowledge bases, tool integrations, and observability out of the box. The core of Astro AI is the astropods.yml spec which is a single file that captures your entire agent topology: the models it uses, the knowledge stores it's backed by, the tools it can call. Version-controlled, reproducible, and GitOps-ready. Three commands take you from idea to production: ast create scaffolds your agent and spec, ast dev runs it locally with hot-reload against live models and tools, and ast push deploys it to the Astro AI registry with tracing, scaling, and versioning included. Astro AI is framework-agnostic, so bring agents built with CrewAI, LangChain, Mastra, or your own stack. Swap model providers without touching code. Run on our managed platform or self-host in your own cloud. Apache 2.0 licensed, open-source to the core. Built for teams shipping real AI agents, not demos.