About Starside Labs
We are a small open source AI lab based in Singapore. We build practical, local-first tools that put people ahead of hype, and we ship every one of them in the open.
What we believe
AI is most valuable exactly where the stakes are highest: your footage, your meaning, your knowledge. Those are the places you can least afford to hand control to someone else's server. So we build tools that run locally where it matters, stay open so anyone can audit them, and measure success by how the work lands with a real person.
Our work is organized into three initiatives. Each one is a long-term thread, and each active project ladders up to one of them.
Core values
Open by default
Permissive licenses, public code, real documentation. If you cannot read it and fork it, we have not finished it.
Local and private
Where the data is sensitive, the computation stays on your machine. Privacy is a default, not a paid tier.
Built for people
We care how the output feels to a human, not just how it scores. Accessibility and clarity over spectacle.
Honest about status
Early access is labeled early access. We tell you what is solid and what is still moving.
Curiosity-led
We work at odd intersections: vision and privacy, neuroscience and translation. The interesting problems live in between.
Community first
Issues, forks and contributions are the point, not a side effect. Build with us.
Three initiatives
Privacy-First AI
Local-first computer vision and analytics. Useful intelligence, zero data exposure.
Petrify →AI & Cognition
Brain-predictive models that measure how machine output lands in a human mind.
Langdrift →Knowledge & Comms
One canonical model, many audience-aware renderings.
Verso & Neurowire →Get in touch
Questions, ideas, or want to contribute?