Applied intelligence
Products where models are the engine, not a feature. Language, vision and voice composed into software that feels obvious.
Sanlabs · Independent AI studio
We design and build AI-native products, autonomous agents and tools of our own — quietly, carefully, and only when they are ready.
Sanlabs is an in-house studio working where artificial intelligence meets software craft. We don't take clients and we don't chase trends. We choose problems worth solving, build the tools ourselves, test them in the real world — and then go further.
In-house. Every line of code, every model pipeline, every pixel.
Products shipped to the App Store — and four more currently in orbit.
External clients. By design. Our only brief is curiosity.
Products where models are the engine, not a feature. Language, vision and voice composed into software that feels obvious.
Systems that plan, use tools and finish real work on their own — with judgement and guardrails designed in from day one.
We study models, markets and human behaviour, prototyping fast to find out what is genuinely worth building next.
Native apps, web, infrastructure. Built end-to-end by a small team that obsesses over the last one percent.
Five products designed, built and released from inside the lab. Each one started as an experiment — and grew up.

One selfie becomes studio portraits, editorial shoots and cinematic reels — across more than 1,500 styles.

The whole Bible as a calm, swipeable feed. One verse per screen, with a plain-English study one tap away.

Describe a mood, paste lyrics or drop in a photo — and get a full song back in seconds, even in your own voice.

Adaptive multi-phase frequencies push water and dust out of iPhone speakers. Plus a tone generator and sound meter.

Two-factor codes and passwords in one zero-knowledge, encrypted vault. Your keys never leave the device.
A glimpse of what is taking shape inside the lab. Names are codenames. Details stay inside until the work speaks for itself.
Agentic workspace for deep, multi-step work
Voice-first companion that remembers context
Evaluation toolkit for models in production
██████████ ████ ███████████
Sanlabs