Press Kit
Resources and assets for media coverage of Cordum.
Company Boilerplate
Cordum is an open-source AI agent orchestration platform that provides governance for autonomous workflows. The platform features a Safety Kernel for policy-before-dispatch evaluation, human-in-the-loop approval gates, and complete audit trails. Cordum is MCP-native (Model Context Protocol) and works with Claude, GPT, and any compatible AI agent. Founded to bridge the gap between AI agent demos and production deployments, Cordum enables enterprises to run autonomous agents safely with deterministic scheduling, policy-as-code, and compliance-ready audit logs.
Key Messages
“The missing governance layer for AI agents in production.”
Company tagline
“Policy-before-dispatch means every agent action is evaluated before execution, not after.”
Technical overview
“Cordum keeps the core boring so you can innovate at the edges.”
Philosophy
What is Cordum?
Cordum is a control plane for AI agents that evaluates every action against configurable policies before execution. Unlike traditional workflow engines, Cordum is built specifically for autonomous AI agents that need governance.
Core Features
- Safety Kernel — Policy-before-dispatch evaluation with ALLOW, DENY, REQUIRE_APPROVAL, or ALLOW_WITH_CONSTRAINTS decisions
- Approval Gates — Human-in-the-loop for high-risk operations, bound to policy snapshot + job hash
- Audit Trail — Append-only timeline for every workflow run with full reproducibility
- MCP Native — Model Context Protocol support for Claude, GPT, and other AI assistants
- Pack System — Installable overlays that extend the platform without core changes
Quick Facts
- Founded
- 2024
- Headquarters
- Remote-first
- License
- BUSL-1.1
- Core Technology
- Go, NATS, Redis