Loading tool…
Policy gate that checks every AI agent tool call before it runs and writes a signed audit trail
Exogram sits between an AI agent and everything it can touch. A local adaptor runs beside the agent, evaluates each proposed tool call against deterministic rules with no model in the decision path, and blocks the ones that would leak customer data, write to a database they should not, or loop on an API and burn budget. Every decision lands in a cryptographic ledger aimed at SOC 2, HIPAA, NIST AI RMF and EU AI Act auditors. It attaches over MCP, a REST API or a CLI and is pitched at LangChain, AutoGen, CrewAI and Claude setups. The edge component is open core; the cloud dashboard and policy distribution are the paid part.
The local adaptor is free and open core; the free tier covers one human seat, one monitored agent and 500 evaluations a month. Single User is $29 a month for up to three agents, Team is $99 a month flat covering five seats and 25 agents, and Enterprise with on-premise or air-gapped deployment is quoted.
Use tool ↗Keeping the model out of the safety decision is the right architectural call — a deterministic gate cannot be talked out of a rule the way an LLM judge can. Open-sourcing the enforcement side and charging for the fleet dashboard is a fair split. What to weigh is that the site's own numbers (0.07ms, 59,946 evaluations, zero failures) are self-reported and no company or country is named anywhere.
Watch out: The free tier caps at 500 evaluations a month on one agent — enough to test the wiring, not enough to leave running.
No reviews yet — be the first to review Exogram.
Reviews are tied to your EffectHub account — one review per tool, so the rating for Exogram reflects real users.
No questions yet — ask the first one about Exogram.
Questions about Exogram are tied to your EffectHub account, so answers can reach you and the section stays free of spam.