An open-source memory layer for autonomous agents that records agent actions into a knowledge graph you can view, search, and configure.
- Storage
- Knowledge graph (e.g., Neo4j) tracking agent actions and entities, paired with vector retrieval.
- Retrieval
- Recursive retrieval: extracts key entities from a query, builds subgraphs to a max depth of 2, and joins them via multi-hop reasoning to reduce latency versus searching the full graph.
- Self-host
- Self-host: moderate
- License
- MIT
- Pricing
- Open source (MIT), free to self-host · Free / OSS
- GitHub stars
- 2,623
- Last release
- 2024-10-22
- Last commit
- 2024-10-18
- First catalogued
- 2026-06-28
Strengths
- Knowledge-graph memory with viewable/searchable agent actions
- Recursive subgraph retrieval with bounded depth for lower latency
- Permissive MIT license
Watch out
- Appears stale: last commit Oct 2024 (~20 months with no activity) — likely unmaintained
- Requires a graph backend (e.g., Neo4j) to self-host
Best for
- Agents that benefit from an inspectable knowledge-graph memory of their actions
Benchmark results
No sourced results yet.
Sources
Last verified 2026-06-28 · updated by manual-refresh