Audio | Tram AI | Documentation

Text-to-Speech

POST /v1/audio/speech

Tạo audio từ text. Response trả về byte stream audio thô.

Request

$ curl https://api.staging.tram.ai.vn/v1/audio/speech \
>   -H "Content-Type: application/json" \
>   -H "Authorization: Bearer $TRAM_AI_API_KEY" \
>   -d '{
>   "model": "~openai/tts-1",
>   "input": "Xin chào!",
>   "voice": "alloy"
> }' --output speech.mp3

Parameters

Parameter	Type	Required	Mô tả
`model`	string	Yes	Model ID
`input`	string	Yes	Text cần chuyển thành giọng nói
`voice`	string	Yes	Giọng nói (alloy, echo, fable, onyx, nova, shimmer)
`response_format`	string	No	Format audio: mp3 (default), opus, aac, flac
`speed`	number	No	Tốc độ (0.25-4.0, default 1.0)

Speech-to-Text

POST /v1/audio/transcriptions

Transcribe audio thành text (multipart/form-data).

Request

$ curl https://api.staging.tram.ai.vn/v1/audio/transcriptions \
>   -H "Authorization: Bearer $TRAM_AI_API_KEY" \
>   -F "file=@audio.mp3" \
>   -F "model=~openai/whisper-1"

Parameters

Parameter	Type	Required	Mô tả
`model`	string	Yes	Model ID
`file`	file	Yes	File audio
`language`	string	No	Ngôn ngữ (ISO 639-1)
`response_format`	string	No	Format output: json (default), text, srt, vtt