使用 memsearch 搜索跨代理 Markdown 内存和对话历史
Give supported coding agents a shared Markdown-first memory layer so they can recall prior decisions, transcripts, and notes across sessions instead of re-deriving context from scratch.
Prerequisites
Supported coding-agent client, local filesystem access, default local Milvus/ONNX setup or another configured embedding provider
Installation
Use the upstream install or setup path that matches your environment:
- uv tool install memsearch # via uv
- pipx install memsearch # via pipx
- pip install memsearch # plain pip
- uv add memsearch # via uv, adds to pyproject.toml
Requirements and caveats from upstream:
- <a href="https://pypi.org/project/memsearch/"><img src="https://img.shields.io/badge/python-%3E%3D3.10-blue?style=flat-square&logo=python&logoColor=white" alt="Python"></a>
- 👥 For Agent Users, install a plugin and get persistent memory with zero effort; for Agent Developers, use the full CLI and [Python API](https://zilliztech.github....
- Beyond ready-to-use plugins, memsearch provides a complete CLI and Python API for building memory into your own agents. Whether you're adding persistent context to a custom agent, building a memory-augmented RAG p...
Basic usage or getting-started notes:
- ### How Plugins Work (Claude Code as example)
- # memsearch command or any of the agent plugins (Claude Code, Codex,
- # OpenClaw, OpenCode), which all shell out to the CLI.
- Source: https://github.com/1991513ccie-png/skills
- Extracted from upstream docs: https://raw.githubusercontent.com/zilliztech/memsearch/HEAD/README.md
Documentation
- https://github.com/1991513ccie-png/skills
