CodeYam CLI & Memory

Mihir Kanzariya, Jarmo Tuisk, and Pär Winzell looked at the AI coding revolution and saw something everyone else missed—these assistants write code at insane speeds but have the memory of a goldfish. You know that feeling when Claude Code forgets critical context three files away? When it rewrites the same thing you just spent hours explaining? That's the problem CodeYam solves. Comprehensive memory management for Claude Code—except it's not some cloud API thing, it's free, runs locally, needs zero registration, and works right now.

Here's how it works: CodeYam watches your coding sessions, catches the decisions that matter and the confusion points that will bite you later, and turns them into rules your agent actually follows next time. You get a dashboard to see every rule, find conflicts, clean up stale references, and track what parts of your codebase have coverage. When your agent suggests something new, you review and approve it. When something's outdated, it gets flagged. It's like giving your AI a notebook that actually pays attention.

The CLI installs in one line, runs in your project root, and they're already working on support for more agents beyond Claude Code. Eight people built this—Mihir, Jarmo, Pär, Dani Rsk, Jared Cosulich, Nadia Eldeib, Amelia Wampler, and Ivo Gospodinov—and they're solving one of the most annoying problems in AI-assisted development without making it complicated. No signup, no cloud, just a local database that makes your agent smarter every time you use it.