GET

/v1/schemas/fal/audio/mirelo-ai/sfx1.6/text-to-audio

Accept:

kind inputactivity audiohash b1f4303b…extracted Aug 29, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `3efadef420ebb4754eda3fa465f8b3aa12d0f277565f0870fd3c44e211d03355` { "fetchedAt": 1787979655, "extractorVersion": "1", "verifiedAt": null }

Request body

7 properties · 1 required · 0 $defs

property	type	description	constraints
text_prompt*	string	Text prompt for the generated audio.	length ≥ 4
seed	integer	Seed for generation. -1 or None for random.	-1 ≤ n ≤ 18446744073709552000 · nullable
num_samples	integer	Number of variations to generate.	1 ≤ n ≤ 4 · default: 2
duration	number	Duration of the generated audio in seconds. The minimum is 0.1s for ordinary sound effects and 1s when ambience=true.	0.1 ≤ n ≤ 60 · default: 10
double_output	boolean	Only used when ambience=true: concatenate the loop with itself for a 2x-length output.	default: false
ambience	boolean	When true, generate a seamlessly loopable ambience tile (room tone, rain, drones, etc.). When false, generate a generic text-conditioned SFX of `duration` seconds.	default: false
upload_audio_format	enum	Audio format for the returned files.	wav \| mp3 \| aac \| flac · default: "wav"

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"fal","endpointId":"mirelo-ai/sfx1.6/text-to-audio","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}