Repository navigation
LoopGrid LangGraph Integration v0.1.0
Initial public release of the native LangGraph integration for LoopGrid's cryptographically verifiable AI decision evidence.
Features
- Native LangGraph decision, guarded tool-execution, and verification nodes.
- Signed decision evidence with SHA-256 commitments and policy provenance.
- Application-controlled policy enforcement and authenticated human-review gates.
- LangGraph checkpoint, interrupt, and resume compatibility.
- Cross-restart replay protection and optional cross-worker dispatch coordination.
- Privacy-aware Core evidence correlation and independent verification support.
Validation
- 89 automated tests passing.
- GitHub CI passing on Ubuntu, Windows, and macOS with Python 3.10, 3.12, and 3.13.
- Successful clean-wheel consumer installation.
- Real LangGraph runtime and persistent SQLite checkpoint testing.
- Successful local LoopGrid Core E2E with signed evidence verification, zero verification failures, and synthetic actions only.
Installation
pip install loopgrid-langgraph==0.1.0Resources
- Repository: https://github.com/loopgridio/loopgrid-langgraph
- LoopGrid: https://loopgrid.io/
- Documentation: https://github.com/loopgridio/loopgrid-langgraph#readme
Security note: The application remains responsible for real-world action execution, authenticated reviewer identity, and deployment-grade cross-worker coordination. The v0.1.0 guarded execution interface is synchronous.
Apache-2.0 licensed.