# MemPalace/mempalace

The best-benchmarked open-source AI memory system. And it's free.

Repository: https://github.com/MemPalace/mempalace
Canonical: https://ross.abutalabs.com/products/mempalace
Homepage: http://mempalaceofficial.com/
Language: Python
License: MIT
License Family: permissive
Topics: ai, chromadb, llm, mcp, memory, python
Last push: 2026-08-26T09:13:28+00:00

## Health v2 (maintenance only)
Score: 81/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 99, longevity 10
- inputs: {"age_days": 151, "days_push": 7, "days_rel": 10, "gap_med": 6, "n_releases_24m": 16}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 58666, forks 7523 (observed 2026-08-28T04:12:18.628787+00:00)

## What it is
MemPalace is a local-first AI memory system that stores conversation history and project files as verbatim text and retrieves them via semantic search using a pluggable vector backend (default ChromaDB). It organizes memories into a hierarchical 'memory palace' structure (wings, rooms, closets, drawers) with a symbolic index, ships as a Python CLI, and requires no API key.

## Use cases
- give my AI assistant long-term memory of past conversations
- search my Claude and ChatGPT chat history locally
- store project files and notes in a searchable memory index
- add persistent memory to Claude Code sessions
- run a local RAG memory store without API keys
- retrieve exact verbatim past conversations for an LLM
- index chat exports from Slack and ChatGPT for semantic search

## When to choose
- you want verbatim, privacy-preserving local memory with no API calls
- you need scoped retrieval over long conversation history (people/projects/topics)
- you want an MCP-integrated memory for Claude Code or similar agents
- you prefer a pluggable vector backend and MIT-licensed open source

## When to avoid
- you need a hosted or multi-user cloud memory service
- you want summarized or compressed memories rather than verbatim storage
- you need native Android/Termux support without a PRoot container
- you need a backend other than vector search without writing a plugin

## Facets
- artifact type: cli-tool
- maturity: active
- function: rag, vector-database, search-engine, mcp, cli, llm-inference
- domain: artificial-intelligence, large-language-models, developer-tools, privacy, self-hosted
- platform: python, cli, windows, cross-platform, self-hosted
- tags: ai-memory, long-term-memory, conversation-history, semantic-search, chromadb, local-first, verbatim-storage, method-of-loci, claude-code, chatgpt-exports, privacy-preserving, no-api-key, retrieval-augmented-generation, ai-agents, linux, macos

## Member repositories
- MemPalace/mempalace (main) score 81

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:18.628787+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 2026-08-29T16:18:19.608366+00:00, confidence not recorded.
  - readme: https://github.com/MemPalace/mempalace (fetched 2026-08-28T04:12:18.628787+00:00, sha 713d262df626)
  - homepage: http://mempalaceofficial.com/ (fetched 2026-08-28T18:01:22.223488+00:00, sha 1a96075ab775)
  - site_page: http://mempalaceofficial.com/guide/getting-started.html (fetched 2026-08-28T18:01:22.232253+00:00, sha 38c0c97b099d)
  - site_page: http://mempalaceofficial.com/guide/getting-started (fetched 2026-08-28T18:01:22.234088+00:00, sha 38c0c97b099d)
  - registry_pypi: https://pypi.org/pypi/mempalace/json (fetched 2026-08-28T18:01:22.235655+00:00, sha a9e8ee6dce1d)
- Data as of 2026-08-30T08:39:29.467469+00:00.