# Hachly

- **Event:** [Gemini Vibe Code Hackathon – London](https://cerebralvalley.ai/e/vibe-code-gemini-london)
- **When:** Sat, Nov 15 at 9:00 AM – 10:00 PM (UTC)
- **Where:** Shoreditch Exchange with Oneder, London, UK
- **Team:** [Eldar Agayev](https://cerebralvalley.ai/u/eldar)
- **GitHub:** https://github.com/eldaragayev/liedetectorgemini
- **Demo video:** https://www.loom.com/share/8072299f2dff4165925e644bcece1065
- **Gallery:** https://cerebralvalley.ai/e/vibe-code-gemini-london/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/vibe-code-gemini-london/hackathon/gallery/28

A "Vibe-powered" AI CHEATER FINDER or you can call it interview engine that integrates multiple Google services—Gemini, Speech-to-Text, Text-to-Speech, and MediaPipe—to generate questions, conduct spoken interviews, and analyze each answer using voice, facial, and linguistic signals. 

The system produces a structured “truthfulness/confidence” score per answer by combining transcripts, timing, prosody, and real-time facial metrics.

Essentially, it's a lie detector that can be used during work interviews or in romantic relationships

Current apps rarely use more than text-only LLM calls and lack easy ways to integrate multiple Google SDKs into real, production-like experiences. This project provides a reusable template that shows how to orchestrate Gemini + audio + vision inside a single Vibe app.

---

Markdown version of https://cerebralvalley.ai/e/vibe-code-gemini-london/hackathon/gallery/28. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt