Skip to main content

/audio/transcriptions

Use this to loadbalance across Azure + OpenAI.

Quick Start​

LiteLLM Python SDK​

from litellm import transcription
import os 

# set api keys 
os.environ["OPENAI_API_KEY"] = ""
audio_file = open("/path/to/audio.mp3", "rb")

response = transcription(model="whisper-1", file=audio_file)

print(f"response: {response}")

LiteLLM Proxy​

Add model to config​

model_list:
- model_name: whisper
  litellm_params:
    model: whisper-1
    api_key: os.environ/OPENAI_API_KEY
  model_info:
    mode: audio_transcription
    
general_settings:
  master_key: sk-1234

Start proxy​

litellm --config /path/to/config.yaml 

# RUNNING on http://0.0.0.0:8000

Test​

curl --location 'http://0.0.0.0:8000/v1/audio/transcriptions' \
--header 'Authorization: Bearer sk-1234' \
--form 'file=@"/Users/krrishdholakia/Downloads/gettysburg.wav"' \
--form 'model="whisper"'

Fallbacks​

LiteLLM supports fallbacks for audio transcription, allowing you to automatically retry failed requests with different models or providers.

Client-Side Fallbacks​

Use fallbacks directly in your code with the LiteLLM SDK:

from litellm import transcription
import os 

os.environ["OPENAI_API_KEY"] = "your-key"
audio_file = open("/path/to/audio.mp3", "rb")

response = transcription(
    model="openai/whisper-1",  # Primary model
    file=audio_file,
    fallbacks=["azure/whisper-deployment"],  # Fallback model
    timeout=30,
    num_retries=1
)

print(f"response: {response}")

Router Fallbacks​

Configure fallbacks using the Router for automatic load balancing:

from litellm import Router
import os

router = Router(
    model_list=[
        {
            "model_name": "whisper-primary",
            "litellm_params": {
                "model": "whisper-1",
                "api_key": os.environ["OPENAI_API_KEY"],
            }
        },
        {
            "model_name": "whisper-backup",
            "litellm_params": {
                "model": "azure/whisper-deployment",
                "api_key": os.environ["AZURE_API_KEY"],
                "api_base": os.environ["AZURE_API_BASE"],
                "api_version": "2024-02-15-preview"
            }
        }
    ],
    fallbacks=[{"whisper-primary": ["whisper-backup"]}],
    num_retries=2
)

# Use with fallbacks
audio_file = open("/path/to/audio.mp3", "rb")
response = await router.atranscription(
    file=audio_file,
    model="whisper-primary"
)

Proxy Fallbacks​

Configure fallbacks in your proxy YAML config:

model_list:
  - model_name: whisper-primary
    litellm_params:
      model: whisper-1
      api_key: os.environ/OPENAI_API_KEY
    model_info:
      mode: audio_transcription
      
  - model_name: whisper-backup
    litellm_params:
      model: azure/whisper-deployment
      api_version: 2024-02-15-preview
      api_base: os.environ/AZURE_API_BASE
      api_key: os.environ/AZURE_API_KEY
    model_info:
      mode: audio_transcription

router_settings:
  fallbacks: [{"whisper-primary": ["whisper-backup"]}]
  num_retries: 2

general_settings:
  master_key: sk-1234

Test Fallbacks​

Test fallback behavior by forcing the primary model to fail:

from litellm import transcription

response = transcription(
    model="openai/whisper-1",
    file=audio_file,
    fallbacks=["azure/whisper-deployment"],
    mock_testing_fallbacks=True  # Force primary to fail
)

Supported Providers​