zilliztech/

memsearch

GitHub Website

MemSearch provides a persistent, unified memory layer for AI coding agents with a Python-based backend, Milvus-backed indexing, and multi-platform plugin support. Latest releases show ongoing maintenance through 0.4.17 (2026-07-31).

2.4kstars

217forks

233issues

MITlicense

2026since

Star historydaily snapshots by VibeCrowd

Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).

Alternatives & relatedmatched by topic overlap

zhayujie/

CowAgent

▲ 135/wk

Open-source super AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-model, multi-channel. Lightweight, extensible, one-line install. (formerly chatgpt-on-wechat)

46k10kPython

thedotmack/

claude-mem

▲ 660/wk

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

90k7.8kJavaScript

bytedance/

deer-flow

▲ 1.2k/wk

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of tasks that could take minutes to hours.

79k11kPython

ruvnet/

ruflo

▲ 629/wk

🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, RAG integration, and native Claude Code / Codex / Hermes and many more Integrated

67k8.0kTypeScript

Leonxlnx/

taste-skill

Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop

72k4.9kJavaScriptgiant

jeecgboot/

JeecgBoot

▲ 83/wk

【低代码迈入v2.0时代，一句话即可生成整个系统】企业级AI低代码平台，一键生成前后端代码甚至整个系统。 AI Skills 一句话画流程、设计表单、生成报表、大屏。内置 AI应用平台涵盖：AI聊天、知识库、流程编排、MCP插件等，兼容主流大模型。引领AI低代码「Skills 生成 → 在线配置 → 代码生成 → 手工合并->AI修改」开发模式，解决 Java 项目 90% 重复工作，提高效率又不失灵活。

47k16kJava

Comparememsearch vs CowAgent memsearch vs claude-mem memsearch vs deer-flow

Reviewgenerated from repository data · Aug 5, 2026

What it is

MemSearch is a persistent, unified memory layer for AI coding agents, backed by Markdown and Milvus, with cross-platform plugin support (Claude Code, OpenClaw, OpenCode, Codex CLI).

How it works

The project uses a Milvus backend as a shadow index and stores memories as Markdown files. It supports multiple embedding providers and configurable Milvus backends, including Milvus Lite, Zilliz Cloud, and a self-hosted Milvus server. It includes features like memory journals (PROJECT.md, USER.md), a skills-from-memory workflow, and a procedural memory layer via installable agent skills.

Getting started

Install and configure across platforms:

Embedding provider defaults to onnx; can switch with:

memsearch config set embedding.provider onnx     # default — local, free
memsearch config set embedding.provider openai   # needs OPENAI_API_KEY
memsearch config set embedding.provider ollama   # local, any model

Milvus backend switch examples:

memsearch config get milvus.uri   # → ~/.memsearch/milvus.db

Milvus Cloud setup example:

memsearch config set milvus.uri "https://in03-xxx.api.gcp-us-west1.zillizcloud.com"
memsearch config set milvus.token "your-api-key"

Self-hosted Milvus (Docker) example:

memsearch config set milvus.uri http://localhost:19530

Configure capture summarization routing and advanced memory maintenance via the config section. See the README for details.

Details and platform-specific setup are provided in the repository documentation linked in the README.

Recent releases

Latest releases:

v0.4.17 v0.4.17 (2026-07-31): Changes include version bumps and fixes for journals and prompt budgets.
v0.4.16 v0.4.16 (2026-07-23): Changes include reducing SessionStart configuration overhead and new configurable index exclusions.
v0.4.15 v0.4.15 (2026-07-20): Surface pending skill candidates and record index health state.
v0.4.14 v0.4.14 (2026-07-13): Limit recent memory previews to daily journals and fix memory-related issues.
v0.4.13 v0.4.13 (2026-07-10): Skip empty sessions in memory previews and scope stale cleanup.

Traction

Stars: 2420 Forks: 217 Open issues: 233

Behind the repo

No startup/company linkage information provided beyond repository and organization in the description.

Caveats

License: MIT Created: 2026-02-09 Last push: 2026-07-31 Language: Python

Agent skills inside · 3

All skills →

memory-recallSearch and recall relevant memories from past sessions via memsearch. Use when the user's question could benefit from historical context, past decisions, debugging notes, previous conversations, or project knowledge -- especially questions like 'what did I decide about X', 'why did we do Y', or 'have I seen this before'. Also use when you see `[memsearch] Memory available` hints injected via SessionStart or UserPromptSubmit. Typical flow: search for 3-5 chunks, expand the most relevant, optionally deep-drill into original transcripts via the anchor format. Skip when the question is purely abouAI & Agents memory-configDiagnose and configure MemSearch memory behavior for the Claude Code plugin. Use when the user asks about MemSearch configuration, plugin summarization, PROJECT.md/USER.md maintenance, memory directories, index health, provider routing, prompt files, or migration/compatibility questions.Databases memory-to-skillTurn workflows from your MemSearch memory into reusable skills. Use when the user asks to make/create/extract/distill a skill from what they just did or from past work, review skill candidates, install a distilled skill, or 'turn this into a skill'. Manages MemSearch procedural-memory candidates under .memsearch/skill-candidates/, not Claude Code's own skills system.Code Review & Quality

SharePost on X LinkedIn