AUDIO WORKSPACE / 01
Words, in sync.
Turn a recording into a precise, word-by-word transcript. Start with your voice or bring an audio clip.
Gnani STT MMS_FA aligner
YOUR SOURCE
Start with audio
or
Drop your audio in hereor browse files on your device
Use your microphone for a quick voice note, or upload a clip instead.
Up to 500 MB Up to 10 minutesWAV, MP3, M4A, WebM
Gnani STT automatically detects the audio language; temporary API-host audio is removed after alignment.
THE RESULT
Your transcript
Your words will land here.
Upload or record something to see your transcript, with every word synced to its moment.
01 Add audio02 Get your transcript
WORD-LEVEL PRECISIONGnani torchaudio MMS_FA
API QUICKSTART
Send audio. Get word timings.
POST an audio file and receive a transcript with start and end times for every word.
1 / REQUEST
Upload your audio
curl --request POST "http://localhost:8000/api/v1/transcriptions" --header "accept: application/json" --form "file=@audio.wav"Send the audio as multipart form data in the file field. Replace the base URL with your API host when deployed.
2 / RESPONSE
JSONWord-level timestamps
{
"transcript": "Hello there.",
"duration_seconds": 2.4,
"words": [
{ "word": "Hello", "start_seconds": 0.18, "end_seconds": 0.92 },
{ "word": "there.", "start_seconds": 1.02, "end_seconds": 1.64 }
]
}Times are in seconds relative to the start of the audio clip. Each word includes its own start and end.