Models and voice
Speech and transcription
Text to speech in Nepali, Hindi and English, and speech to text.
Text to speech: POST /v1/audio/speech
Needs the audio:speech scope. The response is the audio file. Billed by length, only when audio is produced.
| Field | Notes |
|---|---|
| input | Required. The text to speak. |
| language | Optional language code such as ne, hi or en. |
| voice | Optional voice name. |
| speed | Optional. |
| response_format | wav, mp3, opus, aac, flac or pcm. |
curl https://setu.supremeai.services/api/setu/v1/audio/speech \
-H "Authorization: Bearer $SETU_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "input": "नमस्ते! तपाईंलाई कस्तो छ?", "language": "ne", "response_format": "mp3" }' \
--output hello.mp3Speech to text: POST /v1/audio/transcriptions
Needs the audio:transcribe scope. Billed by file size (about one minute per MB).
| Field (multipart) | Notes |
|---|---|
| file | Required. Up to 15 MB. |
| language | Optional language code such as ne, hi or en. |
| response_format | json, text or verbose_json. |
curl https://setu.supremeai.services/api/setu/v1/audio/transcriptions \
-H "Authorization: Bearer $SETU_API_KEY" \
-F file=@voice-note.webm \
-F language=ne \
-F response_format=jsonImages
POST /v1/images/generations is built but switched off for now; calls answer 503 media_disabled.