A2A-native, built for fleets
Every agent is a spec-compliant A2A 1.0 server — agent card, JSON-RPC over /a2a, SSE streaming, tasks/* lifecycle, push. Delegate to other a2a / openai / acp endpoints; run one or orchestrate many.
A small core on LangGraph that grows with git-URL plugins. Run one agent or orchestrate a fleet — from a console, the OpenAI API, or A2A. Local-first, yours to fork.
This site follows the Diátaxis framework:
| Section | Purpose | Start here if you… |
|---|---|---|
| Tutorials | Learning-oriented walkthroughs | Are about to fork protoAgent for the first time |
| How-To Guides | Task-oriented procedures | Need to accomplish a specific change in a fork |
| Reference | Technical descriptions | Need exact details on an API, config key, or extension |
| Explanation | Understanding-oriented discussion | Want to understand why the template is shaped this way |
protoLabsAI/roxy is a filled-in fork built on this template — an autonomous ProtoMaker portfolio manager (persona, A2A skills, project registry). When the docs here don't cover something specific, Roxy is the filled-in example to consult.