Oruk Information

Oruk builds speech foundation models that connect what people say with how they say it. Our API returns transcripts with expressed emotion and speaking-style labels, giving developers acoustic context that text alone misses. Built for teams developing voice products, reviewing customer conversations, and studying spoken language. Try your own audio in the demo, then integrate with our Python or TypeScript SDKs. Our team and advisors bring speech, AI, and NLP research experience from Stanford, Cambridge, and UC Berkeley, alongside engineering and product backgrounds at Meta, Microsoft Azure, and CERN. Backed by a16z speedrun. The first company selected for both a16z speedrun Alpha and speedrun (Alpha 001 / SR007). We publish original experiments, evaluation methods, and limitations so you can examine the evidence before building. Explore our speech-model research, public benchmarks, and open-source Orukeet ASR. Try the demo: https://demo.oruk.ai Build with the API: https://oruk.ai/docs Explore the research: https://oruk.ai/research For integrations and research collaborations: access@oruk.ai