Is Jev actually wired in here, or just mentioned in the README?

Wired in, but optional. Memory search results can be reranked through the TypeSafe API with no local model download, and the repo ships a Chinese/English reranking evaluation for it. It needs a TypeSafe key; everything else runs without one.

Does my conversation history leave the machine?

The memories themselves are Markdown files you can read and edit; the vector store is a derived, rebuildable cache. Embeddings default to a local ONNX bge-m3 model on CPU with no API key, and about 558 MB is downloaded from Hugging Face on first launch.

Is this only for Claude Code?

No. Plugins exist for Claude Code, Codex, DeepSeek Harness, OpenClaw, and OpenCode, all sharing one backend, so a conversation in one agent becomes searchable context in the others.

What will bite me during install?

Codex has to be launched with --yolo so the ONNX model can reach the network. OpenClaw requires explicitly allowing conversation access and prompt injection, then a gateway restart. On OpenCode a background daemon does the capture.

How does recall work under the hood?

Three layers: search, expand, then transcript. Retrieval combines dense vectors, BM25 sparse search, and RRF reranking, SHA-256 content hashing skips unchanged content, and a file watcher auto-indexes new files in real time.

Can I change the embedding provider?

Yes. It defaults to local ONNX bge-m3 on CPU with no API key, and memsearch config set embedding.provider switches to openai, which needs OPENAI_API_KEY, or ollama for any local model. One backend is shared across all the agent plugins.

What are the extra memory layers beyond recall?

A third procedural layer distills repeated workflows into installable agent skills that follow the open Agent Skills standard and stay updated in the background. Optional maintenance tasks keep durable PROJECT.md and USER.md notes current across sessions.

Jev directory

CommunityJev Tools & CLIs

MemSearch

A cross-platform memory layer for AI coding agents: conversations from Claude Code, Codex, DeepSeek Harness, OpenClaw, and OpenCode become Markdown memories indexed in Milvus, with optional Jev reranking through the TypeSafe API.

See highlights Open github.com

Stars: 2.8k
Language: Python
License: MIT
Reading time: 3 min
Published by: Community
Author: zilliztech
Last verified: Sep 2026

zilliztech/memsearch

Organization repository on GitHub

View on GitHub

A persistent, unified memory layer for all your AI agents (e.g. Claude Code, Codex, DSH), backed by Markdown and Milvus.

Forks: 268
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

About this project

Cross-platform semantic memory for AI coding agents. Optional Jev reranking — rerank memory search results with Jev through the TypeSafe API, with no local model download. See configuration and the Chinese/English evaluation. DeepSeek Harness support — MemSearch now brings automatic capture, pre-step memory injection, native skill-based recall, background maintenance, and a read-only memory browser to DeepSeek Harness (DSH).

Quoted from the project's README. Read the full README.

Tagscommunity python agents search reranking claude-code

Works forJev for Code Jev for Tool calling Jev for Ranking Jev for Document processing

Used inJev for Software & DevTools

Highlights

Optional Jev reranking runs through the TypeSafe API with no local model download, and ships with a Chinese/English reranking evaluation.
Markdown is the source of truth and Milvus is a derived, rebuildable index; SHA-256 content hashing skips unchanged files and a watcher auto-indexes.
Recall is three layers — search, expand, then transcript — over dense vectors plus BM25 sparse and RRF reranking.
A third procedural-memory layer distills repeated workflows into installable skills that follow the open Agent Skills standard.
Defaults to local ONNX bge-m3 on CPU with no API key or cost; the roughly 558 MB model downloads from Hugging Face on first launch.

Quickstart

bash

/plugin marketplace add zilliztech/memsearch
/plugin install memsearch

Watch out

MIT-licensed. Plugins exist for Claude Code, Codex, DeepSeek Harness, OpenClaw, and OpenCode; Jev reranking needs a TypeSafe API key, and multi-user setups need a Docker-hosted Milvus.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: Search and tagging on keep.md

X: Jev reranks station suggestions in TrainLCD

YouTube: Jev in the reranking step of a RAG pipeline: why cosine similarity fails and how a steerable reranker compares. A Colab notebook comes with it.

X: LinkedIn bookmarks search

From the Jev guides

Original write-ups on what this entry is for.

Retrieval with Decisions: Filter, Rerank, and Verify

Oct 12, 2026

Retrieval is a sequence of bounded judgments. How a decision model slots into passage filtering, semantic find, reranking, and citation checks, and what stays in deterministic code.

Jev Cookbooks#retrieval#reranking#rag

Read article

Jev at home: voice pipelines, Assist routing, and sentinels

Oct 16, 2026

How Home Assistant conversation agents, macOS voice pipelines, and an advisory sentinel put a fast typed decision in the loop — one call per utterance, thresholds in code, and a daily token budget.

Jev Tools & CLIs#home-assistant#voice#automation

Read article

Search visibility, audited by decisions

Oct 16, 2026

Jev-based SEO tools split the crawl from the call: deterministic crawlers count and parse, while typed judgments rank fixes, place internal links, score search results, and track AI-answer mentions — with costs and caveats attached.

Jev Tools & CLIs#seo#search-visibility#audits

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai

MemSearch

About this project

Highlights

Quickstart

Questions about this entry

Is Jev actually wired in here, or just mentioned in the README?

Does my conversation history leave the machine?

Is this only for Claude Code?

What will bite me during install?

How does recall work under the hood?

Can I change the embedding provider?

What are the extra memory layers beyond recall?

Reactions & coverage

Related terms

More Jev resources like this

From the Jev guides

From the Jev AI model community