GET
/v1/schemas/fish-audio/audio/v1/voice-design
Accept:
| field | recorded source |
|---|---|
| schema | Source linked above. upstream-spec Source trace698642cc0dd247e854fa9711b7d1d26916770ad466e2ae77dc9d983379dc5647{
"fetchedAt": 1794288623,
"extractorVersion": "2",
"verifiedAt": null
} |
Request body
9 properties · 1 required · 0 $defs| property | type | description | constraints |
|---|---|---|---|
| instruction* | string | Voice design prompt. Must contain 1 to 2000 characters. | length ≥ 1 · length ≤ 2000 |
| reference_text | string | Optional text used as reference content for the generated voice. | length ≤ 150 · nullable |
| language | string | Optional BCP-47 language hint, such as `en`, `zh`, or `ja`. | nullable |
| n | integer | Number of voice candidates to generate. | 1 ≤ n ≤ 4 · default: 2 |
| speed | number | Speaking speed multiplier for candidate generation. | 0 < n ≤ 3 · default: 1 |
| num_step | integer | Number of diffusion steps used by the voice-design model. | 1 ≤ n ≤ 128 · default: 32 |
| guidance_scale | number | Classifier-free guidance scale. Higher values follow the prompt more strongly. | n ≥ 0 · default: 2 |
| instruct_guidance_scale | number | Instruction guidance scale for prompt conditioning. | n ≥ 0 · default: 0 |
| seed | integer | Optional deterministic seed for candidate generation. | nullable |
Validate a payload
before you spend tokensshell
$ curl -X POST https://modelschemas.com/v1/validate \
-d '{"provider":"fish-audio","endpointId":"v1/voice-design","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}