The API provides methods to transcribe voice messages.
Schema:
updateTranscribedAudio#84cd5a flags:# pending:flags.0?true peer:Peer msg_id:int transcription_id:long text:string = Update;
messages.transcribedAudio#93752c52 flags:# pending:flags.0?true transcription_id:long text:string = messages.TranscribedAudio;
---functions---
messages.transcribeAudio#269e9a49 peer:InputPeer msg_id:int = messages.TranscribedAudio;
messages.rateTranscribedAudio#7f1d072f peer:InputPeer msg_id:int transcription_id:long good:Bool = Bool;Use messages.transcribeAudio to initiate transcription of a message.
The returned messages.transcribedAudio constructor will have the pending flag set if the transcription is still in progress and the transcribed text contained in text will be updated in future with updateTranscribedAudio updates.
These updates will contain the updated text with the same transcription_id returned in the first messages.transcribedAudio, and the pending flag will be set if the transcription is still in progress.
A transcription can then be rated as good or bad using messages.rateTranscribedAudio.