HELLLOOO!
This is not your preexisting ai call centre. We are fixing the fraud BEFORE calls hit the business. You might think you are too smart to fall for fraudulent calls. But whether it's getting scammed or simply time wasting, fradulent calls affect all people of businesses.
That's why we built Jellyphis, an AI front desk that screens scam calls, so real people in real businesses don't have to.
We built this project using Rust, HTML, Python and JavaScript. ElevenLabs is great and runs the live voice agent. Its webhook tools call a Rust Axum service. Events and call summaries go into SQLite. A plain HTML dashboard allowed us to get live alerts quickly.
The hardest part was getting genuine, quality responses from AI and Elevenlabs in real time. Letting an AI describe requests was what worked, instead of using AI solely as the approval mechanism. The frontend was informative yet clean. We are happy with this. The options avaliable to businesses, and the openness/transparency we showed in our demo was good.
Instead of just a fraud detection and monitoring system, we hope to fully integrate Jellyphis into all existing call agents in the future.
We hope you enjoyed Jellyphis!
Built With
- elevenlabs
- html
- javascript
- python
- rust
Log in or sign up for Devpost to join the conversation.