#347347th of 463 in Skills, plugins and prompts
automem
Graph and vector memory service for AI assistants, exposed over MCP and REST
- Stars
- 827
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Aug 2026
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
AutoMem stores memories in FalkorDB as a graph with typed relationships and keeps an embedding of each in Qdrant. Recall is a hybrid query ranked by a 9-component score, with no generative LLM call. A local MCP bridge connects Claude Desktop, Cursor, Claude Code and Codex, and a remote MCP mode and REST API cover other clients.
Who it is for: Developers who want persistent shared memory across Claude, Cursor and Codex
Strengths
- Recall uses graph and vector lookups with no generative LLM call per query
- 11 authorable relationship types, with multi-hop expansion to surface rationale memories
- Same memory reachable via local MCP bridge, remote MCP over HTTPS, or REST
- Recall still works in degraded mode if Qdrant is down
Weaknesses
- Requires running both FalkorDB and Qdrant alongside the Flask API
- Tags are a hard filter before scoring, so untagged memories can be missed
- Recall does not reliably prefer the newest version of a conflicting fact
- Pre-1.0; entity synthesis and several recall fixes are experimental or opt-in
What it needs
- Docker + Compose
- Needs FalkorDB, Qdrant
- Models: Voyage embeddings, OpenAI embeddings, local embedding provider
- port 8001
Also in Skills, plugins and prompts
See all 463| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |