InterviewLabs

This project is a video-only Smart Interview Analyzer. It:

• Generates interview questions for a given field.
• Accepts a single video containing answers to all questions.
• Extracts audio using ffmpeg, transcribes using local Whisper, and analyzes results with Cohere (if COHERE_API_KEY provided).
• Returns a rating (1-10), timestamped mistakes, and improvement tips.

Requirements (cross-platform)

• Node.js (16+)
• Python 3.8+
• ffmpeg installed and in PATH
• Whisper Python package: pip install -U openai-whisper (or git+https://github.com/openai/whisper.git)
• (Optional) Cohere API key for richer analysis (needed for real ai analysis and question generation create a project.env file containing the api key)

Quick start

1. Unzip project and open terminal in project folder.
2. Install Node deps:

   npm install

3. Install Whisper (Python):

   pip install -U openai-whisper

4. Make sure ffmpeg is installed and available on PATH.
5. Edit project.env and set COHERE_API_KEY if you have one.
6. Run the server:

   npm run dev

7. Open your browser at http://localhost:3000

Notes

• Transcription runs using local Whisper via a Python script transcribe_whisper.py which outputs a JSON transcript with timestamps.
• The server calls that script; ensure your Python environment has Whisper installed.
• If Cohere key is missing, the server falls back to a built-in heuristic analysis.
• ⚠️ Note: The demo hosted on Vercel may not work reliably because the project exceeds the free-tier resource limits. To run it locally, please download the file from releases and follow the setup instructions.