Cordum
Policy and human approval in front of risky AI agent actions, with auditable evidence
Cordum is a policy layer that holds risky AI agent actions for approval before they run. A safety kernel classifies tool calls, shell commands, workflows and production changes, applies policy and asks for human approval where needed, with redacted, hashed audit evidence. It includes a local Edge path for Claude Code, an MCP governance server, a gateway, a workflow engine, a dashboard and a CLI. The core is source-available under BUSL-1.1 and free to self-host; Team and Enterprise licences add capacity and SSO.
Pricing: Free self-hosted core, licensed editions by contract
Choosing a guardrails tool right now?
One click, anonymous. Helps us know our visitors.
Thanks, that helps.
Choosing a guardrails tool right now?
One click, anonymous. Helps us know our visitors.
Thanks, that helps.
Cordum Alternatives
Explore 64 products in the Observability & Analytics category. View all Cordum alternatives.
NeMo Guardrails
NVIDIA toolkit for adding programmable guardrails to LLM conversational apps
LLM Guard
Open-source input and output scanners for LLM apps; the repository was archived in July 2026
Guardrails AI
Open-source framework for adding input and output validators around LLM calls
evalship
Tests, monitors and optimizes the AI calls in your product, from pull request to production
AcruxCore
Open-source LLMOps platform for prompts, gateway, tracing and evaluation
Work on Cordum? Feature it at the top of Observability & Analytics.
Is your product missing?