am looking for a developer who can build a fully custom voice-based AI agent using the OpenAI Realtime API, similar to an interactive voice assistant (not text-based). My company is an HVAC/Plumbing/Electrical home services business, and this AI will act as a virtual specialist that: Speaks in real-time using OpenAI voice Listens and responds naturally Performs lead qualification Performs HVAC/Plumbing repair triage Explains our service plans Collects contact information Creates lead records Books appointments Sends lead data to backend Uses custom tool-calling logic I already have: ✔ Full system prompts ✔ Full conversation scripts ✔ Full flowchart logic ✔ Tool specifications ✔ A backend skeleton (Node.js + Express) ✔ A frontend widget skeleton (HTML/JS/CSS) ✔ Full folder structure delivered You will be building the live, working voice agent on top of this package. Requirements: You MUST have experience with: OpenAI Realtime API WebRTC or WebSockets Streaming audio in/out of browser Node.js backend Tool calling / function calling Integrations with REST endpoints Nice-to-have experience: CRM integration (HubSpot / Sheets / GoHighLevel) Speech barge-in handling Real-time captions Deployed examples of voice AI agents What You Will Build: 1. Frontend (Modify existing files): Connect “ Talk to GFlo” button to OpenAI Realtime API Handle microphone/audio streaming Display live captions (optional) 2. Backend (Using existing Express app): Implement tool endpoints: /api/service-area /api/lead /api/membership-rate /api/appointment And: Create token generation endpoint for Realtime if needed Connect frontend audio → OpenAI → backend tools 3. Voice Agent Logic: System prompt already provided. You ensure correct tool-calling behavior. 4. Output: A fully working, deployed GFlo branded voice agent. Deliverables: ✔ Working frontend widget (connects to Realtime) ✔ Working backend (tool endpoints integrated) ✔ Full audio conversation live in browser ✔ Code uploaded to GitHub or ZIP ✔ Deployment instructions ✔ 7–14 day delivery preferred To Apply, Send: A voice AI or Realtime project you’ve built Your experience with WebRTC Your experience with Node.js tool-calling Estimated price Estimated delivery time