About AgentBreeder

One agent.yaml. Any framework. Any cloud.

AgentBreeder is an open-source platform for building, deploying, and governing enterprise AI agents. It exists so a developer can write one config file, run one command, and ship an agent to AWS, GCP, Azure, or Kubernetes — with RBAC, cost tracking, audit trail, and observability automatic.

Why this exists

Every team building production AI agents reinvents the same plumbing: a deploy pipeline, a registry, RBAC, cost attribution, an audit log, observability. And every framework (LangGraph, CrewAI, OpenAI Agents, Claude SDK, Google ADK) asks you to learn a new container shape, a new entrypoint, a new way to wire secrets.

AgentBreeder collapses that down. Pick your framework. Pick your cloud. Writeagent.yaml. Run agentbreeder deploy. Governance happens as a side effect, not as extra configuration.

The inventor

Rajit Saha

Founder, AgentBreeder · Director of Data Platform, Udemy + Coursera

Rajit has spent 23+ years building distributed systems and data infrastructure across Oracle, IBM, Yahoo, Teradata, VMware, LendingClub, Experian, and Udemy. At the merged Udemy + Coursera company he leads the Data Platform team, where shipping AI agents into production surfaced the exact gap AgentBreeder fills.

AgentBreeder is currently built alongside that role. The open-source project is released under Apache 2.0, with a managed cloud offering (console.agentbreeder.io) coming next.

Where we are today

Open-source CLI + engine — Apache 2.0, installable via pip, Homebrew, or Docker. Multi-cloud, multi-framework, governance built in.
Managed cloud (coming soon) — a hosted control plane at console.agentbreeder.io for teams that want the registry, RBAC, and observability without running the API server themselves.
Status — bootstrapped, no external funding, built in public on GitHub. Contributions and design feedback welcome.

Accessibility

The site is dark-only by design — the whole color system is tuned for dark surfaces and a separate light mode would mean re-tuning every component. If that’s a blocker for you, let us know at hello@agentbreeder.io.

The site respects prefers-reduced-motion and targets WCAG 2.1 AA contrast for all body and meta text. Found an accessibility issue? Open one on GitHub.

Contact

Email hello@agentbreeder.io for partnerships, pilots, press, or just to say hi. For bug reports and feature requests, please open an issue on GitHub.