Zep
Agent memory service that stores facts and their validity windows in a temporal knowledge graph.
0 members list Zep, unchanged over the last 12 weeks, data as of 21 November 2026
Members who list Zep in their Stack, by week.
Members who list Zep
Be the first to add Zep
Nobody on Stackness runs Zep yet. Add it to your stack and you're the first member running it.
About Zep
Zep is a memory service for AI agents that organizes business data, documents, and conversations into a temporal knowledge graph. It stores facts with associated validity windows and retrieves contextual information needed for agent tasks. Developers use Zep to provide agents with governed access to unified business context while reducing token consumption. The service manages context graphs at scale, enforces access policies, and maintains audit logs of queries. Zep is developed by getzep and is licensed under Apache 2.0.
The platform ingests data from multiple business sources and builds context graphs for users, customers, accounts, or domains. It retrieves only the facts required for each task through a filtered policy mechanism, returning results in a single call. Zep includes features for access control, data retention policies, provenance tracking, and maintains sub-200 millisecond retrieval latency regardless of graph size.
Checked 21 November 2026.
Nobody on Stackness runs Zep yet.
Sign up and add itPosts about Zep
- Agent memory format standard: none exists, so keep what must port in the repo
No agent memory format standard exists as of November 2026. Keep what must survive a switch in AGENTS.md and skills, and treat tool memory as a cache.
- Agent memory vs documentation: which slot a memory plugin fills, and which one the repo does
Memory is what an agent noted and a plugin replays. Documentation is what you wrote, reviewed and committed. How each loads, and what to write down.