VOX
One App. All AI Models. Every Voice.
Built for Google Cloud + ElevenLabs Hackathon 2025
AI is fragmented.
Voice is an afterthought.
ChatGPT
1 model, 1 voice, text-first
Claude
1 model, no voice
Gemini
1 model, limited voice
You can't use the best model for each task.
You can't personalize how AI sounds.
That's the gap.
Your AI contacts.
Their voices.
- Multi-model — Gemini, Claude, GPT-4, DeepSeek in one app
- Voice cloning — Hear yourself speak any language
- AI contacts — Each has unique voice, personality, purpose
- Real-time translation — 35+ languages, your voice
Why now?
Voice AI Got Good
ElevenLabs is now indistinguishable from human
Multi-Model Era
No single AI wins everything. Users want choice.
Gen Z is Voice-Native
They grew up with Siri. Voice notes > texting.
Costs Dropping
Voice AI is finally economically viable
2 years ago: Too early → Now: Window open → 2 years later: Too late
"Kids love Voice Mode"
"Kids love voice mode in ChatGPT."
"My kids will never be smarter than AI, but they will grow up vastly more capable than we grew up."
The Thomas the Train Story
A parent put ChatGPT in voice mode with his kid...
"My son thinks ChatGPT is the coolest train-loving person. I can never compete."
What makes us different
Multi-Model
Use Claude for reasoning, Gemini for speed, GPT for coding. All in one interface.
Voice-Per-Contact
Alice's voice = interview mode. Carlos = Spanish mode. Voice triggers instant context.
Group AI Chat
Put Claude, GPT, Gemini in one chat. Watch them debate. Get the best answer.
No one else offers this combination.
What we built
- Multi-model AI (Gemini, Claude, GPT, DeepSeek)
- Voice cloning (ElevenLabs integration)
- Real-time translation (35+ languages)
- Pre-made AI contacts (7 experts)
- Custom contact creation
- Subscription payments (Stripe)
- Cloud sync across devices
Market sizing
Bottom-Up Calculation
| People who pay for AI tools | ~50M globally |
| × Want voice-first experience | ~10% = 5M |
| × Will try a new app | ~20% = 1M |
| × Will convert to paid | ~5% = 50K |
| × Average revenue per user | $15/month |
How we make money
Basic model
Limited contacts
Default presets only
All AI models
Unlimited contacts
Custom contacts
Unlimited everything
Priority access
Early features
To be calculated
ElevenLabs cost per character, LLM API costs, free tier sustainability. This is MVP — we built the product, now we validate the business.
Competitive landscape
| Feature | ChatGPT | Claude | Gemini | Character.AI | VOX |
|---|---|---|---|---|---|
| Multi-model access | — | — | — | — | |
| Voice-per-contact | — | — | — | — | |
| Voice cloning | — | — | — | — | |
| Real-time translation | — | — | Limited | — | |
| Group AI chat | — | — | — | — | Soon |
Window: 12-18 months before big players catch up
What we need to prove
Voice-Per-Contact Aids Memory
Do users recall context faster with unique voices?
Multi-Model Matters
Do users actually switch between models?
Users Will Pay
Free to paid conversion rate
These aren't assumptions. These are experiments we're running.
Why us?
10+ years building products people actually use and return to
What's next
Validate
- → 100 user hypothesis testing
- → Calculate unit economics
- → Voice cloning safety features
- → Security audit
Prove
- → Launch Group AI Chat
- → D7 retention > 25%
- → Free→paid > 5%
- → Optimize API costs
Scale or Pivot
- → If metrics work: Scale
- → If metrics fail: Pivot
- → B2B only if B2C proven
Raising $500K
12 months runway to reach 1,000 paying users
Use of Funds
Milestones
VOX
What we know
- The product works
- The technology exists
- The timing is right
What we'll prove
- Users retain and pay
- Voice-per-contact works
- Unit economics viable
We're not promising the next unicorn.
We're promising to find out if voice-first multi-model AI is a business.
If it is, we're first.