Transcription | FormantAI

Use the transcription API to upload an audio file and poll for the completed transcript.

Submit a job

1 POST /v1/transcription/jobs

$ curl -X POST "https://api.voice.formantai.com/v1/transcription/jobs" \
>   -H "Authorization: Bearer $FORMANT_API_KEY" \
>   -F "audio=@call.wav" \
>   -F "languages=hi-IN,en-IN" \
>   -F "diarize=true" \
>   -F "format=standard"

Form field	Type	Required	Description
`audio`	file	Yes	Audio file to transcribe.
`languages`	string	No	Comma-separated language codes, or `unknown` for auto-detection.
`diarize`	boolean	No	Include speaker segments when supported.
`format`	string	No	`standard` or `code_mixed`.

1 {
2   "id": "job_abc123",
3   "engine": "sarvam",
4   "status": "pending"
5 }

Poll a job

1 GET /v1/transcription/jobs/{job_id}

$ curl "https://api.voice.formantai.com/v1/transcription/jobs/job_abc123" \
>   -H "Authorization: Bearer $FORMANT_API_KEY"

1 {
2   "id": "job_abc123",
3   "engine": "sarvam",
4   "status": "completed",
5   "text": "नमस्ते, मैं राहुल बोल रहा हूं.",
6   "language": "hi-IN",
7   "utterances": [
8     {
9       "speaker": "speaker_0",
10       "text": "नमस्ते, मैं राहुल बोल रहा हूं.",
11       "start": 0.2,
12       "end": 2.8
13     }
14   ]
15 }

Supported language examples

The API supports English and multiple Indic language codes, including:

hi-IN, bn-IN, kn-IN, ml-IN, mr-IN, pa-IN, ta-IN, te-IN, en-IN, gu-IN, as-IN, ur-IN, and unknown.

Format modes

Format	Use when
`standard`	You want normal transcription output.
`code_mixed`	Audio may switch between English and Indian languages.