Automatic project memory for Claude Code. Also works with Cursor and Codex.

What it does: Saves decisions, constraints, bugs and todos from your Claude Code chats into JEVMEM.md, automatically. When you change your mind, the old line is marked superseded, not deleted. Next session, the relevant lines are added to Claude's context.

Install (60 seconds): You need a Type Safe AI key for Jev (an Open AI or Anthropic key is optional). As a Claude Code plugin: `claude plugin marketplace add Avinash-jetwani/jevmem` then `claude plugin install jevmem@jevmem`. With npm: `npm install -g jevmem`, then `cd your-project` and `jevmem init --tool claude`. Init creates JEVMEM.md, jevmem.config.json and a gitignored .jevmem/ folder.

How it decides: Scrub removes common secret shapes, email addresses and card-shaped numbers before the turn leaves your machine. Jev by Type Safe AI answers a fixed set of small questions with probabilities: is there a decision, a rule, a bug? Plain rules over those probabilities decide save or skip. On save, a small LLM (or a deterministic extract) writes one line of at most 200 characters. If the turn replaces an existing memory, that line is tagged `[superseded]` and stays.