mempalace

MemPalace is a local-priority AI memory system that stores speech history as text directly without summarizing it and brings it with semantic search. It uses a structured directory architecture that places people in wings, subjects in rooms, and original content in drawers. It works with ChromaDB as the default backend and supports working locally without calling external API.

ME

MemPalace · AI / LLM / RAG

STARS60.2K
FORKS7.6K
COMMITS2.0K
CONTRIBUTORS141
VERSIONv3.9.0
UPDATEDtoday
PythonAIHafıza SistemiChromaDBMIT

Common commercial products in this category: ChatGPT, Claude, Gemini

What can it do?
  • You can save your speech history as raw text and make semantic calls.
  • You can organize data in a structured directory in the form of wings, rooms and drawers.
  • You can run the system in your local environment using the ChromaDB-based add-on backend structure.
Heads up
  • • High number of open issues (845)
User reviews
  • ↑ PositiveHacker News

    MemPalace, the highest-scoring AI memory system ever benchmarked

    “Many commentators are, mostly fairly, criticizing the repo due to issues raised on benchmarks. This is reasonable, however many are going further to bash the repo likely due to the authors. For me, I see a silver lining. I'll be implementing mempalace for a few small agents to…”

  • ↓ NegativeHacker News

    MemPalace, the highest-scoring AI memory system ever benchmarked

    “The 'store everything vs. synthesize' axis is the interesting design question here. MemPalace bets that retrieval quality is the problem — keep everything verbatim, make it findable. That's a legitimate approach. We went the other direction with Lodestar (launching tomorrow).…”

Verified against README · Oct 26, 2026 · source

My note

Alternative pages

More in: AI / LLM / RAG