mempalace
The best-benchmarked open-source AI memory system. And it's free.
MemPalace is an MCP server that provides AI agents with persistent semantic memory through verbatim conversation storage and retrieval. It stores conversation history as structured text organized into wings (people/projects), rooms (topics), and drawers (content), then retrieves relevant context using local semantic search with 96.6% recall on benchmarks. The server exposes 44 MCP tools covering memory reads/writes, knowledge-graph operations, agent coordination, and cross-wing navigation, enabling AI assistants to maintain long-term context across sessions without cloud dependencies.
Key Features
Use Cases
- 01Preserving Claude Code session context across the 30-day expiration window
- 02Enabling AI assistants to recall past project decisions and architectural discussions
- 03Building multi-agent systems where specialists share knowledge through the palace
- 04Searching conversation history semantically without summarization or cloud APIs
- 05Maintaining long-term memory for coding assistants in Cursor, Windsurf, or Claude Desktop
- 06Coordinating agent handoffs via artifact exchange and logstream events
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
mempalace — FAQ
What is the MemPalace MCP server?+
MemPalace is an MCP server that gives AI assistants persistent semantic memory by storing verbatim conversation history in a structured local index and exposing 44 tools for search, knowledge graphs, and agent coordination. It uses local embeddings and vector search to retrieve relevant context with 96.6% recall without requiring any cloud API.
How do I install the MemPalace MCP server for Claude Code?+
Install with 'uv tool install mempalace' or 'pipx install mempalace', then add a 'mempalace' entry to your MCP settings JSON with 'command': 'mempalace' (or 'uvx mempalace' if using uv without tool install). You can also use the official Docker image mounted to your transcript directories.
Which AI clients work with MemPalace?+
MemPalace works with any MCP-compatible client including Claude Code, Claude Desktop, Cursor IDE, Windsurf, Codex CLI, and Gemini CLI. Auto-save hooks are specifically provided for Claude Code, Codex, and Cursor.
Does MemPalace require API keys or cloud services?+
No API keys or cloud services are required for the core functionality. MemPalace uses local embedding models (all-MiniLM-L6-v2 or embeddinggemma-300m) and runs entirely on your machine. You can optionally configure remote OpenAI-compatible embedding endpoints if desired.
Is MemPalace free?+
Yes, MemPalace is free and open-source under the MIT license.
What are the prerequisites for running MemPalace?+
You need Python 3.9 or higher and approximately 300 MB disk space for the embedding model. The default ChromaDB backend is bundled; other backends like Milvus or pgvector require optional installation with extras like 'mempalace[milvus]'.
How do I install mempalace?+
Open the source repository on GitHub and follow its README. mempalace is a mcp server — MCP Agents Market links you directly to the official repo.
Is mempalace free?+
mempalace is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.