Skip to main content
POST
/
v1
/
stt
/
sarvam
/
saaras:v3
Saaras v3
curl --request POST \
  --url https://api.slng.ai/v1/stt/sarvam/saaras:v3 \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: multipart/form-data' \
  --form audio='@example-file' \
  --form mode=transcribe \
  --form language_code=hi-IN
{
  "transcript": "नमस्ते, आज आप कैसे हैं?"
}

Authorizations

​
Authorization
string
header
required

API key issued by SLNG. Pass as Authorization: Bearer <token>.

Headers

​
X-World-Part-Override
enum<string>

Target world part override. Auto-selected if not provided.

Available options:
ap

Body

multipart/form-data
​
audio
file
required

Audio file (multipart) or base64-encoded audio (JSON).

​
url
string

Publicly accessible audio URL.

Example:

"https://docs.slng.ai/audio/hello.wav"

​
mode
enum<string>
default:transcribe

Output mode.

  • transcribe: Standard transcription in original language.
  • translate: Translate speech to English.
  • verbatim: Exact word-for-word without normalization.
  • translit: Romanization to Latin/Roman script.
  • codemix: Code-mixed (English words in English, Indic in native script).
Available options:
transcribe,
translate,
verbatim,
translit,
codemix
​
language_code
enum<string>
default:unknown

Language of the input audio (BCP-47). Set to 'unknown' for auto-detection.

Available options:
unknown,
hi-IN,
bn-IN,
kn-IN,
ml-IN,
mr-IN,
od-IN,
pa-IN,
ta-IN,
te-IN,
en-IN,
gu-IN,
as-IN,
ur-IN,
ne-IN,
kok-IN,
ks-IN,
sd-IN,
sa-IN,
sat-IN,
mni-IN,
brx-IN,
mai-IN,
doi-IN

Response

Transcription successful.

Sarvam AI STT response format.

​
transcript
string
required

The transcription result.