Add enterprise governance to your LangGraph stateful pipelines. Trust Runtime wraps graph nodes as enforcement checkpoints — so every tool call is validated, every decision is traceable, and your audit trail is production-ready.
The AgentTrust OS LangGraph adapter wraps your graph's tool nodes as enforcement checkpoints. Before any node executes a tool call, Trust Runtime evaluates it against your enforcement contract — auto-approving low-risk actions, escalating edge cases, and blocking policy violations.
npm install @agenttrust/langgraph-adapter or pip install agenttrust-langgraph
A JSON policy file specifying which tool categories are pre-approved, which require human review, and which are blocked outright.
One import. The adapter intercepts tool calls at the node level — no changes to your existing graph structure required.
Trust Audit captures every node execution, tool call outcome, and enforcement decision. Export compliance reports on demand.
Trust Certify stress-tests your LangGraph pipeline before it ships — prompt injection, tool scope, compliance alignment.
Every tool call evaluated before execution. Auto-approve, escalate, or block based on your policy.
Full-trace audit log. SOC 2 / ISO 42001 / GDPR evidence exported in auditor-ready format.
Trust Runtime integrates with LangGraph's checkpoint system — governance-aware state recovery for long-running pipelines.
Start with the Developer plan (free) or book a demo for Enterprise deployment support.
