HeartMuLa models and APIs

About HeartMuLa

HeartMuLa contributes the platform's audio-input model. It is the only creator here whose task starts from a sound file rather than from a prompt, an image or a piece of video.

HeartMuLa Transcribe takes one audio URL and returns the sung lyrics as text. It is multilingual, has no options to tune, and is billed per transcription rather than per second or per token, so the cost of a call is known before you make it. The response is a text artefact, which means it still follows the same task lifecycle as an image or a video: create the task, observe it to a terminal state, then read the result.

Use it when the thing you need out of a recording is words. Because there are no parameters, the only variables that matter are the audio you send and how you handle the returned text, so most of the integration work here is the upload and retention path rather than prompting. If you need sound generated alongside a picture instead, that belongs to the video families whose tasks render audio against the image.