Cutcall

One click and your business never misses a call again. Our AI, powered by Square, answers, chats, and books clients for you—24/7, no tech skills needed. Instant setup, happy customers.

Comment

Inspiration

As a solo developer, I see every day how small business owners struggle to handle phone calls, follow-up messages, and the digital side of their business. For example, my own barber rarely answers calls or replies to WhatsApp.
I wanted to create an AI assistant that could help business owners handle customer enquiries—so they can focus on their real work.

What it does

Cutcall answers calls for businesses, chats with clients, checks the Square calendar, and books appointments—automatically, any time of day.
Setup is just one click: I import all the data from Square, or, if the business doesn’t have an account, I help them create a new Square store with onboarding via an AI assistant in chat or by phone. This removes the barrier of forms and admin screens.

How I built it

  • Frontend: I built a standalone Vite + React app, using Supabase as the backend.
  • Authentication: I used Supabase Auth for full authentication, including email verification and password recovery.
  • API & Logic: Supabase Edge Functions handle Square data imports, AI assistant booking requests, and webhooks from Nango OAuth.
  • Realtime UI: Supabase realtime tables power live onboarding updates, with no need for custom websockets or polling.
  • Integrations:
    • Square: I used Nango for secure, persistent API credentials and future integrations with other business apps.
    • AI Assistant: ElevenLabs conversation agent, injected with business data and tools to fetch available booking times and create bookings. This setup is ready to extend to other services in the future.
    • Telephony: Telnyx handles incoming call routing, recording, audio playback, or connecting to a realtime AI agent. Businesses can keep their current number or get a new one. When redirecting, Cutcall uses the incoming number to load the correct business data for the AI agent.

Challenges I ran into

  • Integrating real-time voice AI with phone systems was much more complex than text—handling call audio streams, latency, and keeping agent responses natural was a major technical challenge.
  • Telephony is highly regulated: Setting up business telephony systems means dealing with strict legal requirements and complex KYC (Know Your Customer) processes. Many providers have multi-step verification and documentation requirements, making it extra hard to quickly onboard and test these kinds of AI-powered solutions.
  • Making one-click setup possible meant hiding a lot of backend complexity, from OAuth security to real-time UI feedback and seamless credential refresh via Nango.
  • My initial telephony provider (Telnyx) blocked my account for KYC a day before hackaton deadline, forcing a rapid pivot to fallback demos using ElevenLabs.
  • Because phone number KYC can take time, I built a “test call” experience in the dashboard so judges and users could try the AI agent without needing a real phone call.

Accomplishments that I'm proud of

  • End-to-end demo: I delivered a working, full-stack demo where an AI assistant answers client questions, checks real service availability, and books appointments in Square—all through natural conversation.
  • One-click onboarding: I created a seamless flow—users connect Square, and their assistant is ready in minutes, with zero technical complexity.
  • Voice AI integration: I integrated ElevenLabs’ conversational agent to enable natural language booking via both chat and phone, not just web forms.
  • Extensible foundation: The system is designed so I can add new services, booking tools, and integrations for growth beyond salons.

What I learned

  • Voice AI is much harder than chat: Real-time phone conversations (especially with external APIs) require thoughtful engineering and prompt design to keep latency low and responses natural.
  • Onboarding is everything: If setup isn’t instant and frictionless, small businesses will drop off. Investing in seamless, real-time onboarding pays off.
  • Third-party APIs are unpredictable: Providers (telephony, auth, payments) may have outages, delays, or unexpected verification steps. Always build fallback/demo modes.

What's next for Cutcall

  • Pilot & Feedback: My next goal is to onboard a dozen businesses and gather real feedback while refining the vision.
  • Ultimate digitalization pack: Cutcall could become the ultimate digitalization solution for small businesses—offering all Square tools plus a 24/7 AI assistant.
  • Growth potential: There are many gaps in the Square ecosystem (like crypto payments). By being the entry point to Square, Cutcall can create referral opportunities and serve as the gateway to new digital services for businesses.

Built With

Share this project:

Updates