Skip to content

Latest commit

 

History

History
96 lines (71 loc) · 2.37 KB

File metadata and controls

96 lines (71 loc) · 2.37 KB

Whisper AI Transcriber Setup

ChatGPT-Level Accuracy (95%+)

This uses OpenAI's Whisper API - the same technology powering ChatGPT voice mode.

Features

  • ✅ 95%+ accuracy (same as ChatGPT voice)
  • ✅ 99 languages including Bulgarian (bg)
  • ✅ Auto language detection
  • ✅ Handles accents, background noise, multiple speakers
  • ✅ Real-time continuous transcription

Installation

pip install openai pyaudio

Or double-click: install_whisper.bat

Get API Key

  1. Go to: https://platform.openai.com/api-keys
  2. Sign up / Log in
  3. Click "Create new secret key"
  4. Copy the key (starts with sk-...)

Cost

  • $0.006 per minute (~$0.36 per hour)
  • Much cheaper than human transcription
  • Pay only for what you use

Usage

  1. Run: python audio_transcriber_whisper.py
  2. Paste your API key
  3. Click "Set Key"
  4. Select microphone
  5. Choose language:
    • auto - Auto-detect (recommended)
    • en - English
    • bg - Bulgarian
    • es - Spanish
    • etc.
  6. Click "Start Listening"

Tips for Best Accuracy

  • Chunk Duration:

    • 3-5 seconds = faster response, good for conversations
    • 7-10 seconds = better accuracy for long sentences
  • Language Setting:

    • Use auto for mixed languages
    • Set specific language (e.g., bg) for better accuracy in that language
  • Microphone:

    • Use a good quality mic (HyperX, Trust USB)
    • Speak clearly at normal pace
    • Reduce background noise

Supported Languages

English, Bulgarian, Spanish, French, German, Italian, Portuguese, Russian, Japanese, Chinese, Arabic, Turkish, Polish, Ukrainian, and 90+ more!

Comparison

Feature Vosk (Free) Google API Whisper AI
Accuracy 70% 85% 95%+
Bulgarian
Offline
Cost Free Free (limited) $0.006/min
Quality Basic Good ChatGPT-level

Troubleshooting

"Invalid API key"

  • Make sure you copied the full key (starts with sk-)
  • Check you have credits in your OpenAI account

"Insufficient quota"

Slow transcription

  • Lower chunk duration to 3-4 seconds
  • Check your internet connection

Poor accuracy

  • Set specific language instead of auto
  • Speak closer to microphone
  • Reduce background noise