About Aristra Labs

Aristra Labs is a small team based in Kuala Lumpur, Malaysia. DF and Ervin started it because they were tired of re-explaining themselves to AI, even tools that claimed to "remember."

The memory problem isn't what you think.

ChatGPT has memory. Claude has memory. Most AI platforms do now. But what they're actually doing is storing facts, floating fragments of text retrieved by vector similarity. "User prefers bullet points." "User is building a SaaS product." It's a filing cabinet that finds things that look alike. It doesn't know what anything means.

It doesn't know that the decision you made last quarter changes how you'd think about the roadmap today. It doesn't know that two ideas you mentioned six weeks apart are actually the same problem. It doesn't know what to let go of, and what to hold onto.

Storing facts isn't memory. It's just a database with a confidence score.

We think real memory is relational. It's understanding how ideas connect, how your thinking evolves, and what context is actually still relevant. Not just retrieving what's similar, but reasoning across what's happened.

That's what Aristra is built to do.

Not a smarter chatbot. Not a better notes app. A cognitive layer that builds a real model of how you think, and gets sharper the longer you use it.

Why we started this

We've been building with AI for years. The capability wasn't the frustration. It was the feeling of starting over, every single time. Brilliant tool, no continuity. You couldn't build anything lasting with something that forgot yesterday.

So we started Aristra. The thesis: the next real step isn't a bigger model. It's a model that actually knows you, not just what you've typed, but how you think.

We're early. The product is in early access. We're building in the open, with the people who use it.

This is Aristra Labs.