Compare speech APIs

Choose by the speech task

Speech recognition, emotion measurement, and voice generation solve different problems. Compare services that return the output you need before comparing a headline score or price.

Transcripts, word timing, and speaker turns

Start with the speech-to-text API and compare Deepgram or AssemblyAI. Check language coverage, streaming status, and transcription errors on your audio.

Emotion and speaking style in recorded audio

Compare the emotion API output with Hume expression measurement. Label sets, selected scores, and transcript sentiment are not interchangeable.

Live transcription with phrase emotion

Run the Realtime preview example and inspect its events. Oruk’s full speaking-style and unified-analysis outputs use separate file endpoints.

Continuous valence and arousal

Choose a dimensional model if you need continuous coordinates. Our valence and arousal guide explains how to evaluate those predictions and why Oruk’s categorical scores are a different output.