SocialCue

The voice-first co-pilot that whispers confidence into your ear. Mastering the art of small talk for international students and technical founders. Makes your life easier

Comment

SocialCue — Voice-First Social Intelligence

Inspiration

Networking is the unwritten requirement of success.

But for international students navigating language barriers, or technical founders who are more comfortable with code than cocktail hours, small talk can feel like a terrifying obstacle.

We realized something important:

The barrier isn’t competence — it’s social latency.

That split-second hesitation. The pause while searching for the right word. The anxiety of an awkward silence.

That moment alone can kill a connection before it ever starts.

Most existing tools focus on grading your speech after you fail.

We wanted something different.

We wanted to help people win in the moment.

We believe the future of soft skills is Voice-First. Not a chatbot. Not a scorecard.

But a real-time conversational safety net — one that levels the playing field for non-native speakers and introverts alike.

What is SocialCue?

SocialCue is a real-time, voice-first AI agent built on LiveKit.

It acts as an invisible coach during simulated networking scenarios, continuously listening and intervening instantly to prevent social stumbling blocks — before they happen.

Think of it as:

A best friend in the room, quietly helping you sound confident.

Core Innovations

① Whisper Push — Real-Time Rescue

Our flagship feature.

When SocialCue detects:

  • A freeze
  • A prolonged hesitation
  • A non-native speaker struggling to find the right English phrase

…it intervenes.

The AI literally whispers a relevant topic, phrase, or transition directly into the user’s ear.

✔ No interruption
✔ No embarrassment
✔ No awkward silence

Just-in-time help — exactly when you need it.

② Audio Nudge — Confidence Correction

Confidence is heard in silence, not noise.

When users rely on filler words like:

  • “um”
  • “uh”
  • “like”

SocialCue emits a subtle, non-intrusive audio chime.

This gently trains the brain to:

Pause instead of panic

Over time, this builds:

  • Authority
  • Clarity
  • Presence

All without breaking conversational flow.

③ Game Tape — Post-Interaction Analysis

Professional athletes watch game tape.
Professional networkers should too.

After each session, users receive:

  • A breakdown of conversational flow
  • Highlights of moments they commanded the room
  • Identification of moments where a Whisper Push was needed

It’s reflection — without judgment.

Technical Challenges We Faced

Orchestrating the Race Conditions

Running two AI models simultaneously introduced serious synchronization challenges.

Timing wasn’t just important — it was everything.

The “Double Talk” Problem

We had to ensure:

  • The OpenAI agent never responded to the 11Labs whisper
  • Coaching audio stayed invisible to the AI partner

This required strict audio isolation:

  • Separate tracks
  • Precise routing
  • Zero leakage

Only the human hears the coach.
The conversation stays authentic.

The Vision

SocialCue isn’t about replacing human interaction.

It’s about removing friction —
so intelligence, personality, and opportunity can finally speak for themselves.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Built With

Share this project:

Updates