A stealthy, context-aware copilot that bridges Google Meet captions with Sarvam AI to provide real-time technical interview assistance.
This project uses the Microphone Button as a precise physical trigger to control the AI. It does not record constantly. Here is the exact flow:
| ➡️ | While you are muted, the script silently buffers the interviewer's words from the captions. It waits for you. | |
| ➡️ |
This is the Magic Moment. When you click "Unmute" to answer, the script detects the state change (OFF → ON). It assumes: "The user is speaking, so the interviewer must be finished." It immediately sends the buffered question to the AI. |
|
| ➡️ | The AI sends a flash answer to your Console instantly. Simultaneously, the script deletes the old captions from the Google Meet screen. This ensures you don't send duplicate text and keeps your screen clean. |
This project is a "Stealth Bridge" that connects Google Meet to a local Python server. It uses Sarvam AI (or Gemini) to process questions in real-time. Instead of generic answers, it injects your specific Resume and Persona to generate tailored, technical responses displayed discreetly in your Developer Console.
This project is optimized for Sarvam AI. It is highly recommended due to its speed and generous developer benefits.
- 💰 ₹1000 Free Credits: Generous tier for new developers.
- ⚡ Low Latency: Faster response times for English/Indic contexts.
- 🚀 Unlimited Chat: Available on specific dashboard interfaces.
👉 Get API Key: https://dashboard.sarvam.ai/
|
We use # Install required packages
uv pip install flask flask-cors openai python-dotenv |
Create a # Choose your provider:
SARVAM_API_KEY=sk-xxxxxx
GEMINI_API_KEY=xxxxxx |
| File Name | Description |
|---|---|
resume.txt |
Paste your full resume here. The AI matches answers to your experience. |
system_prompt.txt |
Define the persona. (e.g., "You are a Senior React Dev.") |
|
Run the backend script using uv run main.py |
1. Open Google Meet and turn on Captions (CC). |
- 🖱️ Zero-Latency Trigger: Uses the mute button state to trigger AI instantly. No waiting for silence.
- 🧠 Resume Injection: The AI "reads" your resume before answering.
- 🧹 Auto-Wipe: Deletes used captions from the screen to prevent confusion.
- 🔌 Multi-Model: Supports Sarvam AI, Google Gemini, and Perplexity.
PLEASE READ CAREFULLY BEFORE USING THIS SOFTWARE:
This software is provided for educational and research purposes only (e.g., to demonstrate real-time API integration and browser automation).
- No Liability: The developer/author of this repository assumes ZERO responsibility for any consequences arising from the use of this tool. You agree that you are solely responsible for your actions.
- Professional Consequences: Using this tool during an actual employment interview, exam, or professional assessment may be considered fraud, cheating, or a violation of contract/Terms of Service. This can lead to immediate disqualification, job loss, or legal action against you by the interviewing company.
- Legal Action: By downloading or using this code, you agree to indemnify and hold the developer harmless from any legal claims, lawsuits, or demands arising from your usage of this software.
USE AT YOUR OWN RISK.
This project is licensed under the MIT License.