Haldir is an open-source governance layer for AI agents, adding identity, secrets management, audit, and policy enforcement capabilities. Designed for organizations running AI agents in production, Haldir provides scoped agent sessions, encrypted vault storage for secrets, immutable hash-chained audit trails, human-in-the-loop approval workflows, and a proxy that enforces policy on every model context protocol (MCP) tool call. It works with popular agent frameworks like LangChain, CrewAI, AutoGen, and the Vercel AI SDK. Haldir can be self-hosted (ideal for regulated or air-gapped environments) or accessed via a managed cloud service—both with full feature parity.
Visit Haldir's official website for product details and getting started.
Comprehensive API reference and integration guides for Haldir.
Join discussions, ask questions, and get support from other users and developers.
Insights and updates on AI governance and best practices for using Haldir.