Welcome to the Idun blog

L'essentiel: Idun Group is launching this blog to share how we take AI agents from notebook prototype to production. Expect product updates on Idun Engine (open-source runtime, GPL 3.0) and Idun Platform (governance control plane), technical deep dives on guardrails, observability, memory and MCP tooling, and best-practice patterns drawn from real client deployments at LVMH, Schneider, L'Oréal and Leeto. Everything we ship for agents is open source and self-hostable.

We are excited to launch the Idun Platform blog - your go-to source for updates, technical deep dives, and best practices around deploying AI agents to production.

What to expect

Here is what we will be covering:

- Product updates - New features, integrations, and improvements to the platform - Technical deep dives - Architecture decisions, performance tuning, and implementation guides - Best practices - Patterns for guardrails, observability, memory management, and MCP tool governance - Community highlights - Showcasing what the community is building with Idun

Why we built Idun

Getting an AI agent working in a notebook is one thing. Running it reliably in production - with authentication, observability, guardrails, and memory - is another challenge entirely.

Idun bridges that gap. You bring your LangChain, LangGraph, ADK, or n8n agent, and Idun wraps it into a production-grade service with:

- Standardized APIs (AG-UI, CopilotKit, REST + Streaming) - Built-in observability via OpenTelemetry - 15+ guardrails out of the box - Conversation memory with pluggable backends - SSO and RBAC for enterprise deployments

All self-hosted, all open source.

Stay tuned

We have a lot of content planned. Star us on GitHub and check back here for regular updates.