Contact person: Business Client

Phone:Show

Email:Show

Location: Montreal, Canada

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I want to build a web-based AI assistant that users can speak to naturally and receive instant, accurate replies. The core requirement is solid voice recognition so spoken input is captured flawlessly and passed to the language model that powers the assistant’s reasoning.

Here’s the workflow I have in mind: a visitor clicks a microphone button on the site, speaks a question or request, the assistant transcribes the audio in real time, thinks through the best answer, then responds in text (and, if practical, playable audio). High-quality natural language understanding is essential, but voice capture and transcription are the first priority.

I’m open to whichever stack or APIs you feel are most robust—Web Speech API, Whisper, Deepgram, or similar for the voice layer, and ChatGPT, Claude, or a custom fine-tuned model for the dialogue engine—so long as the final product runs smoothly in modern browsers without forcing users to install anything.

I am also looking for a retail AI specialist that can create the voice AI for calls, in bound and outbound.

Deliverables I expect:
• A responsive web page that hosts the assistant
• Integrated voice input with reliable transcription accuracy
• Conversational back-end that returns coherent, context-aware answers
• Clear setup instructions so I can deploy on my own server

If you’ve built something like this before, show me a quick demo or code snippet so I can see the voice flow in action." (client-provided description)