I'm looking for Gemini Live API voice chatbot and an example html page that interfaces to it. The browser accesses the microphone and streams the audio to Gemini Speech to Text and into a Gemini Agent then to Gemini Text to Speech and back to the browser, so the user on the webpage can have a voice conversation with the voice chatbot. The Gemini agent needs to have 3 parts: a Gemini LLM, gemini-2.5-flash-preview-native-audio-dialog, that is having the conversation with a system prompt about being a project intake agent, a Gemini LLM, gemini-2.0-flash, that is the parsing agent that extracts relevant information into a structured format, a Gemini LLM, gemini-2.0-flash, that is the evaluation agent to determine when the project is fully understood and the conversation is complete, and a Gemini LLM, gemini-2.0-flash, summarization agent that summarizes the conversation after completion. https://ai.google.dev/gemini-api/docs/live
Deliverables:
1. A fully working HTML/JS front-end that demonstrates microphone access and streaming audio to Gemini and playing audio received form Gemini
2. Python backend/API logic to orchestrate the speech to text, agents, manage session state, and text to speech
3. Ability to demonstrate a voice conversation from the webpage with the Gemini agent