We're building persistent memory for AI.

AI assistants are powerful, but they treat every conversation like the first. Decisions you made last week, the way you like your code formatted, the project you've been shaping for months — none of it survives the gap between sessions, let alone the gap between tools.

Vilix exists to close that gap. We believe AI should remember across conversations, across projects, and across providers. Memory is what turns an assistant into a collaborator.

The most useful work happens over weeks and months — not in single sessions. Real projects accumulate context: prior decisions, design constraints, evolving goals, personal preferences. When that context disappears, every new conversation starts from zero.

Persistent memory turns AI from a stateless query engine into something that grows with you. The longer you work with it, the more useful it becomes — because it actually knows you and your work.

Memory is intimate. The decisions, drafts, and half-formed ideas you give to your AI tools belong to you. Vilix is being designed so that you choose where memory lives — local, cloud, or hybrid — and which tools can read or write to it.

We are committed to a model where you can audit, edit, and erase what Vilix remembers. No black boxes. No surprises.

Vilix starts with the most painful version of the problem: switching between ChatGPT, Claude, and code-focused assistants like Cursor. Over time we want to be the memory layer for any AI surface you use — including ones that don't exist yet.

We're an early team building in the open with an early access community. If that resonates, join the waitlist.