moorcheh-ai / memanto

Local-first persistent memory for AI agents that requires no vector database, API keys, or cloud backend.

活跃维护 MIT Python Tracked
1.9k 638 1 天前
CIDockerPyPI
ai-agents memory moorcheh memanto agent-memory crewai langchain llm-memory long-term-memory rag semantic-memory stateful-ai

星标趋势

数据积累中,暂无足够数据生成趋势图

AI 分析

项目摘要

Memanto is a persistent, local-first memory system for AI agents that provides long-term memory capabilities for tools like Claude Code, Cursor, Codex, and 14+ other agents. Built on an information-theoretic search engine, it runs entirely on-device with no API keys, vector databases, or cloud backends required.

为什么值得关注

Memanto differentiates itself by offering agent memory that runs 100% locally without vector databases or cloud dependencies, addressing privacy and ownership concerns while integrating with the most popular AI coding tools. Its rapid adoption (1,678 stars, 51 contributors, 30 releases in 6 months) signals strong product-market fit in the growing agent memory space.

优势

  • 100% local execution with no API keys, vector databases, or cloud backends required
  • Broad integration with 14+ AI agents including Claude Code, Cursor, and Codex
  • Active development with 30 releases in 6 months and 51 contributors
  • Strong documentation ecosystem with dedicated docs site, Discord community, and setup videos

局限性

  • Relatively new project (created March 2026) with limited production track record
  • Faces established competition from projects like Mem0, Zep, and Letta/MemGPT

使用场景

  • Persistent memory for AI coding assistants across sessions
  • Long-term knowledge retention for AI agent workflows
  • Local-first RAG-like capabilities without cloud dependencies
目标用户: Developers using AI coding assistants who want persistent, private memory without cloud dependencies, and AI agent developers building stateful applications
学习曲线:
分析模型:LongCat-2.0 | 分析时间:1 个月前