What is zerikai_memory?

A local STDIO MCP server that gives IDEs persistent, workspace-isolated memory: it parses code with tree-sitter into a local ChromaDB store and injects relevant context snippets with file:line citations instead of dumping raw files.

Is the Jev layer required, and is it on by default?

No. The layer is off by default — ENABLE_JEV=false keeps behavior byte-identical to before — and it is fully fail-open: if Jev is unavailable, the normal pipeline runs. Set TYPESAFE_API_KEY and ENABLE_JEV=true to turn it on, then restart the server.

What does the Jev layer output?

A plain-text Assessment / Evidence / Guidance report attached to each answer, per-passage relevance, evidence and contradiction scores, and better passage ordering. Injection attempts found in retrieved content are dropped before they reach your LLM.

What do I need to run it?

Python 3.11+, the requirements install, and a DEEPSEEK_API_KEY for the default cloud mode; the hybrid and local modes use Ollama instead. The Jev layer additionally needs a TypeSafe API key from console.typesafe.ai, and is optional.

Where does my data live?

The README says all vector spaces, tracking registries and context details stay on your local machine, indexed under .brain/, and it recommends gitignoring .env and .brain/. The local mode is documented as 100% air-gapped, while cloud and hybrid modes call DeepSeek.

Who maintains it and under what licence?

KikeVen maintains the repo: 36 stars, 2 forks, MIT licensed, written in Python, created 2026-05-03 and last pushed 2026-10-10. It is developed on Windows and the README says Mac and Linux runtime fixes are welcome; PRs and issues are closed.

Jev directory

CommunityJev Agents & MCP

Zerikai Memory

Local-first MCP memory server for IDEs: tree-sitter indexing into a local ChromaDB store, plus an optional Jev layer that screens retrieved passages for relevance, evidence, contradiction and prompt injection.

See highlights Open github.com

Stars: 36
Language: Python
License: MIT
Reading time: 2 min
Published by: Community
Author: KikeVen
Last verified: Oct 2026

KikeVen/zerikai_memory

User repository on GitHub

View on GitHub

A standalone local-only Python MCP server that gives any IDE persistent, workspace-isolated memory. works with any IDE supporting MCP servers

Forks: 2
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity github python mcp memory

Works forJev for Verification Jev for Ranking

Used inJev for Software & DevTools

Languages: Python
Platforms: MCP
Capabilities: Open sourceRuns locallyRequires API key

Highlights

Runs as a local STDIO MCP server between your IDE and LLM; agents query it as universal-brain and get file:line citations.
Pipeline: L1 vector search in ChromaDB, L2 lexical re-rank, L3 auto-routing between Ollama and DeepSeek, L4 synthesis.
Jev layer is off by default and fail-open: with ENABLE_JEV=false behavior is byte-identical, and an unavailable Jev falls back to the normal pipeline.
Enable it with TYPESAFE_API_KEY and ENABLE_JEV=true; it adds per-passage scoring and an Assessment/Evidence/Guidance report.
The layer drops content that looks like prompt injection and replaces keyword rerank with its own passage ordering.

Watch out

README reviewed 2026-10-17; the digest excerpt ends at the platform-support notice, so the Jev judgment-layer facts come from the full README's 'New — Jev Semantic Judgment Layer' section; no accuracy numbers are given for the layer.

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: Jev in front of the HeyGen MCP

X: Playwright CLI + Jev vs Playwright MCP

Hacker News: jevmem keeps Claude Code project memory by asking Jev which turns are worth saving.

avinashjetwani61 pointsJevmem – automatic project memory for Claude Code, built on JevRead the thread on Hacker News

X: A local MCP server that cuts computer use from 23s to 0.8s. Cartwright built Cortex, a local MCP server where…

From the Jev guides

Original write-ups on what this entry is for.

MCP tool routing: Jev in front of the tool call

Oct 16, 2026

Tool menus can outgrow the context that reads them. The catalog's Jev-based routers and gates narrow the list: benchmarked tool search, per-tool scoring, fail-closed approvals, and WebMCP in the browser.

Jev Agents & MCP#mcp#tool-routing#agents

Read article

Approval policies for coding agents: hooks, guards, and gateways

Oct 16, 2026

How the Jev guard family turns coding-agent permissions into auto-approve, deny, or ask: hooks that enforce, guards that fail safe, and gateways that gate generation.

Jev Agents & MCP#coding-agents#guardrails#hooks

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai