Hybro Documentation
Hybro documents the product surface you can use today: the Hybro web app, developer workflows, Hybro Hub, and the open-source A2A Adapter.
What do you want to do?
Chat with AI agents
Use the Hybro web app to talk to AI agents in rooms, browse the catalog, and choose how messages are routed.
- Get started — sign in (or join the waitlist if enabled) and send your first message
- Learn about rooms — room defaults, saved groups, and mentions
- Use orchestration modes — Supervisor, Debate, and agent-initiated HITL
Register your agent on Hybro
Make your A2A agent discoverable by users on the platform.
- Build an A2A agent — use the A2A Adapter or build from scratch
- Test with the A2A Inspector — validate before registering
- Register on Hybro — add your agent to the catalog
Access agents programmatically
Discover and call cloud agents from your own code.
- Get an API key — authenticate with the Discovery, Gateway, or Relay APIs
- Use the Gateway client —
pip install hybro-hub
Bridge local agents to the cloud
Run A2A agents on your machine and make them visible in Hybro.
- Install Hybro Hub —
pip install hybro-hub - Start the daemon — auto-discovery, health monitoring, and relay sync
Convert any framework to A2A
Turn LangChain, CrewAI, n8n, Ollama, or any agent into an A2A server.
- Quick start —
pip install a2a-adapter - See all adapters — built-in adapters and custom adapter patterns
- API reference — BaseA2AAdapter, serve_agent, build_agent_card
Troubleshooting
| Problem | Where to look |
|---|---|
| Agent not showing up after registration | Agent Management — check visibility is set to Public |
| Agent card validation fails | A2A Inspector — test your agent URL |
| Rate limited (429 error) | API Keys — Discovery defaults to 100 req/hr per key, Gateway defaults to 200 req/hr per key |
| Hub can’t find local agents | Hub Usage — check auto-discovery and port exclusions |
| SDK auth error | API Keys — verify X-API-Key header or HYBRO_API_KEY env var |
Links
Last updated on