langgenie.ai

Getting started

Quickstart

Make your first Odia transcription request in under five minutes.

1. Get an API key

Sign up, open Dashboard → API Keys and click Create new key. Copy the key — it’s only shown once.

2. Transcribe Odia speech

Send any Odia audio file to the speech-to-text endpoint:

import requests

# Odia speech → Odia text
response = requests.post(
    "https://api.langgenie.ai/v1/speech-to-text",
    headers={"api-subscription-key": "YOUR_API_KEY"},
    files={"file": open("sample.wav", "rb")},
    data={"model": "langgenie-od-od-v1", "with_timestamps": "true"},
)

print(response.json()["transcript"])

3. Or translate it to English

Same request shape, different endpoint and model:

import requests

# Odia speech → English text
response = requests.post(
    "https://api.langgenie.ai/v1/speech-translate",
    headers={"api-subscription-key": "YOUR_API_KEY"},
    files={"file": open("sample.wav", "rb")},
    data={"model": "langgenie-od-en-v1", "with_timestamps": "true"},
)

print(response.json()["transcript"])

4. Read the response

200 OK
{
  "request_id": "req_8f2c1e9a4b7d3a10",
  "model": "langgenie-od-od-v1",
  "transcript": "ନମସ୍କାର, ମୁଁ ଆଜି ମୋ ଜମିର ପଟ୍ଟା ବିଷୟରେ ଜାଣିବାକୁ ଚାହୁଁଛି।",
  "source_language": "od-IN",
  "target_language": "od-IN",
  "duration_seconds": 4.2,
  "segments": [
    {
      "start": 0,
      "end": 4.2,
      "text": "…"
    }
  ]
}
Tip: 16 kHz mono WAV gives the best balance of accuracy and upload size. Longer files are supported up to 25 MB per request.
PreviousIntroductionNext Authentication