TensorKit

TensorKit is our AI kits company: small, focused tools for voice, chat and agentic workflows that you can read in an afternoon and put in a product. We opened the organisation in February 2025.

One question runs through all of it: what does it take to put real-time AI, whether speech, agents or generated code, inside an actual product rather than a demo notebook? Every repo in the org is an answer to that at a different stage of maturity.

What is in there

• OpenBeam: real-time Bible verse detection for live sermons, in the browser. The most-starred work in the org.
• Candy: a specification language for stateful backends. Write actors, flows and policies once; generate idiomatic Go, Rust, TypeScript or Python.
• Yugo: a GPU-accelerated interface for the Unitree Go2 robot dog: phone as controller and voice wand, cloud LLM and vision as the mind. A hackathon build.
• Hezra: a voice-powered financial assistant prototype, an OpenAI Realtime voice client over a Go server handling transfers, invoicing and virtual cards.

How we work

Everything here is an experiment. We build to find out what holds up in a real product, and we keep the work public because the code and the design notes may be useful to someone else running the same experiment, not because these are supported products. Some experiments graduate into things people use every week. Most earn their keep by answering a question and then stopping.

That is the same pattern we have run since the research lane of 2023, with one difference: these ship as tools other people install, rather than previews we demo.

If something in the org is useful to you, issues and PRs are welcome on the individual repos.