GET

/v1/schemas/fal/audio/fal-ai/maya/batch

Accept:

kind inputactivity audiohash e3e4d150…extracted Oct 9, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. upstream-spec Source trace `0d5ce29d7951ed0b70d2920a12228520666edb72508f38c9099e60b8127ea73b` { "fetchedAt": 1791522002, "extractorVersion": "1", "verifiedAt": null }

Request body

9 properties · 2 required · 0 $defs

property	type	description	constraints
texts*	string[]	List of texts to synthesize into speech. You can embed emotion tags in each text using the format <emotion_name>.	items ≥ 1 · items ≤ 100
prompts*	string[]	List of voice descriptions for each text. Must match the length of texts list. Each describes the voice/character attributes.	items ≥ 1 · items ≤ 100
output_format	enum	Output audio format for all generated speech files	wav \| mp3 · default: "wav"
sample_rate	enum	Output audio sample rate for all generations. 48 kHz provides higher quality, 24 kHz is faster.	48 kHz \| 24 kHz · default: "48 kHz"
max_tokens	integer	Maximum SNAC tokens per generation.	28 ≤ n ≤ 4000 · default: 2000
repetition_penalty	number	Repetition penalty for all generations.	1 ≤ n ≤ 2 · default: 1.1
top_p	number	Nucleus sampling parameter for all generations.	0 ≤ n ≤ 1 · default: 0.9
enable_safety_checker	boolean	Authorized callers may set false to request relaxed screening.	default: true
temperature	number	Sampling temperature for all generations.	0 ≤ n ≤ 2 · default: 0.4

Validate a payload

before you spend tokens

shell

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