Contact person: Business Client

Phone:Show

Email:Show

Location: Birmingham, United Kingdom

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I need to embed a ChatGPT-powered voice assistant into my existing health website so visitors can talk to the site instead of typing. The conversational flow already lives on the site as on-screen prompts; your job is to connect those exact prompts to a voice interface rather than rewriting or expanding them.

The assistant must comfortably handle two kinds of interactions: guiding users through the health assessments on the site and offering general health tips when asked. When a user speaks, their request should be transcribed, sent to the same ChatGPT prompt logic the text interface uses, and the spoken reply should arrive instantly in natural-sounding speech. If the user’s environment blocks audio, the system should fall back to the original text exchange.

You are free to choose the best implementation method: Web Speech API, custom JavaScript, or an OpenAI Whisper + TTS pipeline; the goal is a smooth, browser-based experience that works on mobile and desktop without forcing extra downloads. Whatever approach you pick, keep the code modular so I can drop future prompts into the same flow easily.

Deliverables
• Fully functional voice assistant embedded on my site, matching the behaviour of current text prompts
• Clean, documented code and configuration files (front-end script and any server hooks)
• A brief setup guide so I can replicate the installation on staging and production

Once everything is in place I’ll test the assistant by running through our full assessment journey and requesting random health tips; the work is considered complete when every spoken exchange mirrors the existing text experience and no console errors appear." (client-provided description)