About Open Multi-Agent.
An MIT-licensed TypeScript framework for multi-agent orchestration — and the people who maintain it.
What this is
Open Multi-Agent (OMA) is a TypeScript-native framework for running teams of AI agents. You hand it a goal or an explicit task DAG; a coordinator decomposes the work, fans agents out in parallel, and returns a traced, resumable run. It runs in your environment — cloud providers, local models, or fully offline.
The runtime ships as @open-multi-agent/core on npm under the MIT license. Development happens in the open on GitHub: issues, releases, and the full commit history are public.
Who maintains it
OMA is authored and maintained by Jack Chen, founder of 元定义科技 (YuanASI), with contributions from the wider community — 53 contributors and 6,788 GitHub stars at the time this page was built.
The project is not backed by a foundation or a venture-funded company product team. Direction is set in public through GitHub issues and releases.
How it is funded
This site carries no advertising, no sponsored placements, and no data-driven monetization. The framework is free to use under MIT and always will be.
YuanASI, a separate commercial entity, offers paid delivery, integration, and support for teams that want OMA implemented on a deadline. That is the funding model — the open-source project itself stays uncommercial.
Reaching us
For bugs, feature requests, and technical questions, open an issue or a discussion on GitHub — that is the primary channel and answers stay searchable for everyone.
For commercial delivery or consulting, see yuanasi.com, or email [email protected] directly.
Last updated: