Audio models
115 models. Compare text-to-speech and speech-to-text models — voice generation and transcription run in opposite directions and are separate catalogue categories.
Audio, on Infery.ai
Audio splits into two catalogue categories. audio_stt is speech-to-text: it transcribes an audio file into text. audio_tts is named for text-to-speech, and about half of its models do exactly that — turn a written script into a voice — but the other half take an existing audio or video clip as input instead, generating sound effects, foley timed to a video, or extending and filling an existing clip. That audio-to-audio work sits inside the same catalogue category as voice generation, not a separate one.
One prompt, 4 models. Credits are what each run actually cost.
It is a catalogue, not one model
Within audio_tts, primary input is the real split: a text-input model turns a script into a voice, usually billed per character; an audio- or video-input model instead transforms existing sound, usually billed per image or per second. audio_stt models differ in maximum input length and language coverage, listed on each model's schema. Narrating a script and transcribing a recording are opposite jobs done by different models — audio_tts alone does not do both.
All 115 audio models
Filter by provider, price and capability. Open one for its rates, accepted inputs and schema.
115 of 115 models
Open the Studio with a audio model already picked
Free trial credits, no card. Filter above, or start from a blank chat and pick as you go.