Echo: The Voice-First Social Network

Inspiration

Social media glues us to screens. We wanted the community depth of Reddit and the raw authenticity of BeReal, but completely hands-free. Echo is designed to help you stay connected while walking, cooking, or driving; without ever looking at a phone.

What it does

Echo is a voice-only social platform that turns your social feed into a personalized daily podcast.

• Daily Briefing: Instead of doom-scrolling, an AI host narrates your friends' updates and interesting forum discussions.
• Voice Forum: Ask questions and answer others using just your voice.
• Interactive 3D: A "living" interface featuring a reactive Audio Ring and Friends Graph that react to your speech in real-time.

How we built it

We combined a visually interesting 3d frontend with a Python FastAPI backend:

• Frontend: Built with React and Tailwind css. We used custom shaders to create the audio-reactive visualizers.
• Backend: FastAPI (Python) and MongoDB manage the data flow.
• AI Engine:
  • Google Gemini: The "brain" that categorizes voice queries, summarizes daily activities, and determines user intent.
  • ElevenLabs: The "voice" that provides ultra-low latency Speech-to-Text and natural Text-to-Speech.

Challenges we faced

• Latency: Making the voice interaction feel instant and conversational required optimizing the pipeline between ElevenLabs and our backend.
• 3D Integration: Seamlessly blending the HTML UI with the 3D canvas so users could interact with the "Friends Graph" naturally.

Accomplishments that we're proud of

• The "Daily Summary": It genuinely feels like listening to a radio station dedicated to your friend group.
• Reactive Visuals: The way the interface breathes and distorts with the user's volume creates a deeply immersive experience.

Built With

• elevenlabs
• fastapi
• gemini-api
• mongodb
• python
• react
• tailwind