thedotmack/claude-mem
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More observed · 2026-08-28
Health v2 · maintenance only
80/100
- Activity 99
- Release rhythm 87
- Longevity 26
How is this computed?
round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.
- gap_med: 0.0
- age_days: 367
- days_rel: 7
- days_push: 7
- n_releases_24m: 309
Adoption not part of the score
91957 stars · 8085 forks observed · 2026-08-28
What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded
claude-mem is an open-source persistent memory engine for AI coding agents such as Claude Code, Codex CLI, Gemini CLI, Cursor, and other MCP-capable tools. It captures structured observations of what agents do during sessions, compresses them with AI into a local-first SQLite database with a Chroma vector index, and injects relevant context back into future sessions via hooks, slash commands, and a bundled MCP server, with an optional CMEM Cloud sync layer.
Use cases
- give my AI coding agent long-term memory across sessions
- stop re-explaining project context to Claude Code every session
- remember decisions and dead ends from previous agent sessions
- sync one agent memory across multiple machines and editors
- search past agent work semantically instead of re-reading the codebase
- share a single memory layer across Claude Code, Cursor, Codex and Gemini
- reduce token costs by reusing compressed agent context
When to choose
- You use Claude Code or another MCP-aware coding agent and want sessions to pick up where the last one left off
- You want local-first memory stored in SQLite on your own machine with privacy stripping and opt-in cloud sync
- You work across several agents, IDEs, or machines and want one shared memory endpoint
- You want out-of-band observation capture that does not block or slow down your agent
When to avoid
- You need a general-purpose RAG pipeline over arbitrary documents rather than memory of agent sessions and interactions
- Your environment cannot run Node 20+ or the Chroma vector stack (including Windows without WSL/Git Bash in older versions)
- You need a turnkey hosted multi-tenant memory service with no local component
- Your agent or editor has no MCP support and you are unwilling to use hooks or CLI integration
Facets
plugin · maturity active
rag mcp vector-database database search-engine artificial-intelligence large-language-models developer-tools cross-platform self-hosted cli agent-memory claude-code mcp-server long-term-memory context-compression session-memory embeddings chromadb sqlite local-first slash-commands hooks cloud-sync memory-engine ai-agents retrieval-augmented-generation nodejs docker
7 sources
- readme: https://github.com/thedotmack/claude-mem · fetched 2026-08-28 · c7c8261fd210
- homepage: https://claude-mem.ai · fetched 2026-08-28 · 7deaa16c4e11
- site_page: https://cmem.ai/features · fetched 2026-08-28 · 7377d6312172
- site_page: https://cmem.ai/about · fetched 2026-08-28 · e7441adeb316
- site_page: https://cmem.ai/integrations · fetched 2026-08-28 · 8f6a70e3a037
- site_page: https://cmem.ai/pricing · fetched 2026-08-28 · 037bd87fd693
- site_page: https://cmem.ai/changelog · fetched 2026-08-28 · c43073637d52
Member repositories
| Repository | Role | Health v2 |
|---|---|---|
| thedotmack/claude-mem | main | 80 |
For agents
markdown · JSON · MCP: product_card(name="thedotmack/claude-mem")
Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem