Documentation Index
Fetch the complete documentation index at: https://mintlify.com/phpaisdk/fal/llms.txt
Use this file to discover all available pages before exploring further.
FalTranscriptionModel sends audio to {baseUrl}/{modelId} (defaulting to https://fal.run/{modelId}) as either a remote URL or a base64-encoded data URI, and returns a TranscriptionResponse containing the transcript text. The provider inspects the audio source automatically — if the source is a URL it is forwarded directly; if it is inline data it is encoded as a data URI — so both input types use the same audio_url field in the request body.
Transcribe from URL
Transcribe from a Data URI
When the audio source is inline base64 data (a data URI), the SDK encodes it appropriately and sends it as the sameaudio_url field. No change is required in the way you call run() — the provider handles the difference transparently:
Request Parameters
The following fields are sent as JSON to the fal.ai endpoint:| Field | Type | Description |
|---|---|---|
audio_url | string | The remote URL or the base64 data URI of the audio to transcribe. |
$request->providerOptionsFor('fal') are merged into the body, allowing you to pass model-specific parameters such as language hints or timestamps settings.
Response
transcribe() returns a TranscriptionResponse with a transcript property of type TranscriptData:
text(string) — the full transcript text, taken from thetextfield of the fal.ai response body.
Transcription is synchronous —
FalTranscriptionModel performs a single POST request and returns the result immediately with no polling loop. The transcriptionPollTimeoutMs property exists in FalOptions but is not used by the current FalTranscriptionModel implementation. See Configuration Options for the full list of available options.