EDN, the Episodic Dialogue Network, is provenance-governed memory middleware. It gives any model persistent, governed recall, with every fact typed, sourced, and eligible by design.
AI systems, agents and applications depend on memory to maintain continuity and act reliably. When context, source and verification are lost, generated content, assumptions and hallucinations can be stored and reused as facts - creating silent errors that persist, spread and influence future decisions.
EDN, the Episodic Dialogue Network, sits between your systems and any model. It captures interactions and documents, types every record by where it came from, and assembles a bounded, governed memory context for each turn. The model stays yours. The memory, its sources, and the rules for what may be recalled stay with EDN.
Many inputs, one governed memory, any model. EDN is infrastructure, not a chat feature.
EDN runs behind a live chat and a governance console today. Ask across sessions and watch it recall with sources attached. Open the memory dashboard and see provenance, integrity, and abstention in real time. Powered by EDN Memory Engine.
[MEMORY CONTEXT] - EU data retained 24 months. (DOCUMENT_VERIFIED) - Deleted within 30 days after. (DOCUMENT_VERIFIED) [END MEMORY CONTEXT]
EDN is a working, fully functional MVP with a sealed evaluation, a provisional patent, and a closed-source core. We are plain about what is live and what is next.