Oruk is a speech API that captures both what people say and how they say it. Developers add transcription, emotion scores, and speaking-style analysis to their apps through one API.
Oruk: Humanizing Speech AI
Oruk helps developers understand both the words and the way they are spoken. Through one speech API, you can add transcription, emotion scores, and speaking-style analysis to an application.
Key Features
Transcription and Vocal-Expression Analysis: For recorded speech.
Emotion and Speaking-Style Scores: Timed results throughout the audio.
Speaker Labels: Available on supported models to help review conversations.
Realtime Preview: Live transcription in 32 locales with phrase-level emotion.
Developer Tools: Python and TypeScript SDKs, plus a hosted MCP server for coding agents.
Orukeet Model: A downloadable model for local transcription in 25 languages, separate from the hosted API.
Use Cases
Call and Meeting Analysis: Add vocal context to conversation review.
Voice Products: Bring emotion and style signals into voice apps.
Interview Review: Analyze interview recordings.
Features and language coverage vary by model. Emotion scores describe vocal expression and are not calibrated probabilities, so evaluate results on audio representative of your application.
Pricing
Monthly plans: Hobby ($9), Builder ($49), and Production ($199), with Enterprise options available. Standard seven-day API trials require a card and continue as the selected paid plan unless canceled before the trial ends.