Overview
Performs speech recognition using Groq’s ultra-fast Whisper API. Groq provides extremely fast inference speeds for Whisper models with high accuracy.
Method Signature
recognize_groq(
audio_data: AudioData,
model: Literal["whisper-large-v3-turbo", "whisper-large-v3"] = "whisper-large-v3-turbo",
language: str | None = None,
prompt: str | None = None,
response_format: str = "json",
temperature: float | None = None
) -> str
Parameters
The audio data to recognize. Must be an AudioData instance.
model
str
default:"whisper-large-v3-turbo"
Groq Whisper model to use:
"whisper-large-v3-turbo" - Faster, optimized model (recommended)"whisper-large-v3" - Full large model for maximum accuracy
Input language as an ISO-639-1 code (e.g., "en", "es", "fr", "de").Specifying the language improves accuracy and reduces latency. If not specified, the model will auto-detect.
Optional text to guide the model’s style or continue a previous audio segment. The prompt should match the audio language.Useful for:
- Specifying spelling of uncommon words
- Providing context about the audio content
- Maintaining consistency across multiple segments
Format of the response. Default is "json".
temperature
float | None
default:"None"
Sampling temperature between 0 and 1. Higher values make output more random, lower values make it more focused and deterministic.
Return Value
The transcribed text from the audio.
Exceptions
Raised if the speech is unintelligible or the API returns an empty transcription.
Raised if the API request fails, the API key is invalid, or there is a network error.
Raised if the groq package is not installed.
Setup
Installation
Install the Groq package:
pip install SpeechRecognition[groq]
API Key
- Visit Groq Console
- Go to API Keys menu
- Generate a new API key
- Set the environment variable:
export GROQ_API_KEY="your-api-key-here"
The groq library will raise a groq.GroqError if the GROQ_API_KEY environment variable is not set.
Examples
Basic Usage
import speech_recognition as sr
import os
# Set API key (or use environment variable)
os.environ["GROQ_API_KEY"] = "your-groq-api-key"
r = sr.Recognizer()
# From microphone
with sr.Microphone() as source:
print("Say something!")
audio = r.listen(source)
try:
text = r.recognize_groq(audio)
print(f"Groq Whisper: {text}")
except sr.UnknownValueError:
print("Could not understand audio")
except sr.RequestError as e:
print(f"API error: {e}")
With Specific Language
import speech_recognition as sr
r = sr.Recognizer()
with sr.Microphone() as source:
audio = r.listen(source)
# Specify language for better accuracy
text = r.recognize_groq(audio, language="es") # Spanish
print(f"Spanish transcription: {text}")
From Audio File
import speech_recognition as sr
r = sr.Recognizer()
with sr.AudioFile('audio.wav') as source:
audio = r.record(source)
text = r.recognize_groq(audio)
print(f"Transcription: {text}")
With Prompt for Context
import speech_recognition as sr
r = sr.Recognizer()
with sr.Microphone() as source:
audio = r.listen(source)
# Provide context to improve accuracy
prompt = "This audio discusses machine learning and artificial intelligence."
text = r.recognize_groq(audio, prompt=prompt)
print(f"Transcription: {text}")
Using Different Models
import speech_recognition as sr
r = sr.Recognizer()
with sr.Microphone() as source:
audio = r.listen(source)
# Use turbo model (faster, recommended)
text_turbo = r.recognize_groq(audio, model="whisper-large-v3-turbo")
# Use full large model (maximum accuracy)
text_large = r.recognize_groq(audio, model="whisper-large-v3")
print(f"Turbo: {text_turbo}")
print(f"Large: {text_large}")
- whisper-large-v3-turbo: Optimized for speed while maintaining high accuracy
- whisper-large-v3: Full model with maximum accuracy
Groq’s inference is significantly faster than standard Whisper API implementations, making it ideal for real-time applications.
Language Support
Supports 99 languages including:
- English (en)
- Spanish (es)
- French (fr)
- German (de)
- Italian (it)
- Portuguese (pt)
- Dutch (nl)
- Russian (ru)
- Chinese (zh)
- Japanese (ja)
- Korean (ko)
- And many more…
See Groq Speech-to-Text documentation for the complete list.
Best Practices
Specify the language explicitly for better accuracy and faster processing.
Use whisper-large-v3-turbo for most applications - it provides excellent accuracy with much faster inference.
Groq requires an API key. The service has generous free tier limits but requires account setup.
External Resources
