mempalace

mempalace

An open-source, local-first AI memory system supporting multiple storage backends.

MCPDevOpen source
Type
MCP
Transport
stdio
Open source
Yes
GitHub Stars
★ 56.2k
Source
mcp-github

Overview

MemPalace is an open-source, local-first AI memory system designed for storing and retrieving conversation history. It stores data in raw text form and retrieves it using semantic search. The system does not summarize, extract, or rewrite content. The indexing structure is organized into wings, rooms, and drawers, enabling scoped searches. By default, ChromaDB is used as the storage backend, but it supports pluggable alternatives such as SQLite, Qdrant, and pgvector. Ideal for developers needing to manage large volumes of conversation data within a local environment.

Capabilities

  • Local-first AI memory system
  • Supports multiple storage backends
  • Structured indexing system
  • Semantic search functionality

Use cases

Managing AI conversation historyData retrieval in local environmentsOrganizing data across multiple projects and topics

Setup

Requires: Python环境API Key(可选)
Recommended installation methods include `uv tool install mempalace` or `pipx install mempalace`. Alternatively, use `pip install mempalace` within a virtual environment.

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

How to choose a storage backend?

You can select different storage backends via the command-line argument `--backend`, such as ChromaDB, SQLite, etc.

Does it support external services?

Yes, it supports external services like Qdrant and pgvector, which require corresponding environment variables to be configured.

Related skills