GET

/v1/schemas/byteplus/audio/tts/create?kind=output

Accept:

kind outputactivity audiohash 32570d15…extracted Aug 15, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. probe-verified Source trace `ebffeaf934197130d6b4cc6534a33b6b04d51a5a71ce2403b2bcf8c7e2cb627f` { "fetchedAt": 1786770726, "extractorVersion": "1", "verifiedAt": "2026-08-12" }

Response body

5 properties · 0 required · 2 $defs

property	type	description
audio	string	Base64-encoded audio in the requested format.
duration	number	Length of the delivered audio in SECONDS (float), after speech_rate is applied.
original_duration	number	Length in SECONDS before rate adjustment — the billing basis, capped at 120.
url	string	Temporary download URL for the same audio, signed with an x-expires deadline.
subtitle	object	Present when audio_config.enable_subtitle was set. Word timings are NESTED inside each sentence — there is no top-level words array — and the block repeats the full text alongside them.

Types

2 $defs

TTSSubtitleSentenceobject

One timed sentence, carrying its own word-level breakdown. Times are MILLISECONDS.

text	string
start_time	number
end_time	number
words	TTSSubtitleWord[]

TTSSubtitleWordobject

One timed word. Times are MILLISECONDS — unlike the response's duration fields, which are seconds. Note the tokenizer emits whitespace as its own zero-or-near-zero-width entry.

text	string
start_time	number
end_time	number

Validate a payload

before you spend tokens

shell

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