# LycheeMem/LycheeMem

Lightweight Long-Term Memory for LLM Agents.

Repository: https://github.com/LycheeMem/LycheeMem
Canonical: https://ross.abutalabs.com/products/lycheemem
Homepage: https://lycheemem.github.io
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agent-memory, llm-memory, openclaw, openclaw-plugin, lychee, mcp, mcp-server, skill, memory, ai, ai-memory, ai-memory-system, memory-system, hermes
Last push: 2026-08-06T14:54:11+00:00

## Health v2 (maintenance only)
Score: 58/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 96, release rhythm 35, longevity 11
- inputs: {"age_days": 163, "days_push": 27, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1155, forks 16 (observed 2026-08-28T04:03:47.834280+00:00)

## What it is
LycheeMem is a lightweight long-term memory framework for LLM agents, combining working session context, a compact semantic memory store (SQLite FTS5 + LanceDB), and a procedural skill layer with HyDE retrieval. It integrates via a Python API, an HTTP MCP server, and native plugins for OpenClaw, Claude Code, and Hermes.

## Use cases
- give my llm agent long-term memory
- remember conversations across chatbot sessions
- add persistent memory to claude code
- store and retrieve reusable skills for ai agents
- run an mcp memory server
- semantic memory store without neo4j
- compress conversation history for agents

## When to choose
- you need lightweight long-term memory for LLM agents without heavy infrastructure like Neo4j
- you want MCP or plugin integration with agent runtimes like OpenClaw or Claude Code
- you want both semantic knowledge and procedural skill memory in one framework

## When to avoid
- you need a fully managed or hosted memory service
- you require graph-based memory with complex entity relations
- your stack is not Python-based and you cannot use MCP

## Facets
- artifact type: library
- maturity: active
- function: rag, vector-database, mcp, plugin-system, search-engine
- domain: artificial-intelligence, large-language-models, developer-tools
- platform: python, self-hosted, cross-platform
- tags: agent-memory, llm-memory, mcp-server, openclaw-plugin, lancedb, sqlite-fts5, semantic-memory, procedural-memory, hyde-retrieval, claude-code-plugin, ai-agents, retrieval-augmented-generation

## Member repositories
- LycheeMem/LycheeMem (main) score 58

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:47.834280+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:32:11.018060+00:00, confidence not recorded.
  - readme: https://github.com/LycheeMem/LycheeMem (fetched 2026-08-28T04:03:47.834280+00:00, sha ce525235a314)
  - homepage: https://lycheemem.github.io (fetched 2026-08-29T12:37:23.174689+00:00, sha 88229880dcb8)
- Data as of 2026-08-30T08:39:29.467469+00:00.