KodLabs is a product and technology lab exploring useful software built around local AI.
We work on efficient models, device-native software, and tools that help people learn, research, create, and build with less friction.
Our areas of interest include:
- Local AI and on-device model execution
- Small, focused models for practical tasks
- AI workflows across phones, tablets, and computers
- Learning and research tools
- Developer tools and local runtimes
- Reliable software that works close to the user
We build through focused experiments: start with a real problem, test the smallest useful system, measure it on real devices, and improve it from there.
- Whyboard - A workspace for writing, learning, and research
- InferPeer - Local coordination for AI workloads across trusted devices
- ModelFit - Understanding which models fit which hardware
- Kernlet - Isolated local execution for software workloads
Our goal is simple:
Build capable software that runs closer to the people who use it.