Orchestrator Integrations
pyhall works alongside your existing orchestrator — not instead of it. All governance decisions flow through the Hall Server HTTP API: POST /api/route on http://localhost:8765.
Your orchestrator keeps orchestrating. pyhall decides who executes what, under what policy, with cryptographic proof.
Supported Integrations
| Integration | Language | Guide |
|---|---|---|
| LangGraph | Python | LangGraph guide → |
| CrewAI | Python | CrewAI guide → |
| AutoGen | Python | AutoGen guide → |
| LlamaIndex | Python | HTTP API guide → |
| HTTP API (any language) | Any | HTTP API guide → |
How It Works
All patterns use the same Hall Server HTTP API. The integration is a single HTTP call before you dispatch a task:
POST http://localhost:8765/api/routeIf allowed is true, proceed. If false, the capability is blocked by policy — check deny_reason.
Setup
- Install pyhall:
pip install pyhall-wcp - Start Hall Monitor (or Hall Server directly):
pyhall hall start - Point your orchestrator at
http://localhost:8765
That’s it. No SDK changes to your orchestrator required — it’s plain HTTP.
See the Hall API reference → for full endpoint documentation.