Inspiration
Interacting with AI is becoming a commodity. In order to stand out, voice agents need personality and a sense of humor is a great step towards that.
What it does
Using voice and video, converses with the user in a humorous way
How we built it
Livekit for voice pipeline Rime for TTS with customization Google Gemini as LLM engine
Challenges we ran into
Compatibility between types in livekit and gemini Pronounciation generation for custom SSML from Rime
Accomplishments that we're proud of
Working demo
What we learned
video capture and processing in livekit Rime details on custom pronounciations
What's next for Atarino
Integrate into our startup's product
Built With
- deepgram
- gemini
- livekit
- rime
Log in or sign up for Devpost to join the conversation.