Audio transcription
Convert uploaded recordings into searchable text.
AI Tools · Function
experimental_transcribe
experimental_transcribe generates a transcript from an audio file using a transcription model. It accepts audio data or a URL and returns the completed transcript as text.
How it works
experimental_transcribe is imported as transcribe and called with a transcription model plus audio content. It returns a result whose text field contains the complete transcript; provider-specific configuration can be supplied through providerOptions.
Syntax
experimental_transcribe({ model, audio, providerOptions? }) → Promise<{ text: string; ... }>Parameters
| Name | Type | Description |
|---|---|---|
model | TranscriptionModelV2 | The transcription model to use. |
audio | DataContent (string | Uint8Array | ArrayBuffer | Buffer) | URL | The audio file to generate the transcript from. |
providerOptions | Record<string, Record<string, JSONValue>> | Provider- or model-specific settings. |
Returns
An object containing text, the complete transcribed text from the audio input.
Key points
- Use the function to convert an audio file into text.
- Pass a transcription model through the
modelproperty. - Provide audio as supported data content or a URL through
audio. - Read the returned
textproperty for the full transcript. - Use
providerOptionsfor provider-specific transcription settings.
Common mistakes
- The API is experimental and may change.
- URL inputs are downloaded automatically and are subject to a default 2 GiB size limit.
Example: Transcribe an audio file
import { experimental_transcribe as transcribe } from 'ai';
import { openai } from '@ai-sdk/openai';
import { readFile } from 'fs/promises';
const { text } = await transcribe({
model: openai.transcription('whisper-1'),
audio: await readFile('audio.mp3'),
});
console.log(text);Reads an audio file, sends it to an OpenAI transcription model, and prints the returned transcript.
When to use it
Captions and meeting notes
Create captions or notes from prerecorded audio.
Customized transcription
Use provider-specific transcription controls.
Related
transcribeTranscriptionModelV2providerOptionsexperimental_streamTranscribe