GetRouterDocs
  • English

    • Audio

    GetRouter exposes three OpenAI-compatible audio operations:

    OperationEndpoint
    TranscriptionPOST /v1/audio/transcriptions
    TranslationPOST /v1/audio/translations
    Text to speechPOST /v1/audio/speech

    Use a model that supports the requested audio operation.

    Transcribe audio

    Upload the source file as multipart form data:

    curl https://getrouter.ai/v1/audio/transcriptions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "[email protected]" \
  -F "model=MODEL_ID" \
  -F "response_format=json"

    The JSON response contains the transcript in text.

    Optional transcription fields in the API schema include:

    • language — ISO-639-1 language code
    • prompt — context or vocabulary hint
    • response_format — json, text, srt, verbose_json, or vtt
    • temperature
    • timestamp_granularities — word or segment

    Support for optional fields depends on the selected model.

    Translate audio

    curl https://getrouter.ai/v1/audio/translations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "[email protected]" \
  -F "model=MODEL_ID" \
  -F "response_format=json"

    The response uses the same basic text shape as transcription.

    Generate speech

    curl https://getrouter.ai/v1/audio/speech \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "input": "GetRouter connects your application to multiple AI models.",
    "voice": "alloy",
    "response_format": "mp3"
  }' \
  --output speech.mp3

    The speech request schema includes:

    FieldDescription
    modelRequired speech model ID
    inputRequired text to synthesize
    voiceRequired voice supported by the model
    response_formatmp3, opus, aac, flac, wav, or pcm
    speedPlayback speed for compatible models

    Python transcription example

    from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://getrouter.ai/v1",
)

with open("recording.mp3", "rb") as audio_file:
    transcript = client.audio.transcriptions.create(
        model="MODEL_ID",
        file=audio_file,
    )

print(transcript.text)
    Tip

    Audio uploads and generated files can be large. Set client timeouts appropriately and stream file data instead of loading unnecessary copies into memory.

    Billing and retries

    Audio billing can depend on the model, input size, duration, or generated output. Check Pricing. Before retrying, determine whether the server completed the request and only the file download failed.

    Next steps