An assistant that stays.

AI has become quite capable. Many people use it to write, explain, plan, calculate, or compare. And yet, working with it often feels exhausting. You keep explaining the same things, lose threads, and thoughts — and entire projects — fall apart into isolated chats. Nothing really builds on what came before. That’s not your fault. And it’s not the AI’s fault either. What’s missing is simple: memory.

What if it could remember? Not just the last conversation, but how you work. What matters to you. Which rules apply. What was decided yesterday. What’s coming up next week. And whatever else you’d like to keep track of. This is where MetaMemoryWorks comes in.

MetaMemoryWorks gives AI memory — outside the model, in plain, ordinary files. Instructions can be stored persistently. Results, decisions, and notes are written down. And you can explicitly define what matters to you. This turns a forgetful text generator into an assistant that accompanies you over time — one you can continuously adapt to your needs, ideas, problems, and ways of thinking.

What does that mean in practice? You no longer work with one AI, but with assistants: for writing, for training, for nutrition, for learning. Not because MetaMemoryWorks tries to solve everything at once, but because the architecture allows you to separate concerns cleanly — while still letting them work together. You start small and gradually shape exactly the assistants you actually need.

What you need to get started: access to an LLM (for example via an existing AI service like ChatGPT or Claude). That’s it. MetaMemoryWorks itself is immediately usable and free for private, non-commercial use. No subscription. No lock-in. No “just trust us.”

If you like, the next step is simple: try a module — and see what it feels like when AI suddenly starts remembering things.