Run locally

1. Install dependencies

npm install

2. Start the dev server (auto-reloads on changes)

npm run dev

Dev server runs on http://localhost:3000 (configurable in vite.config.ts).

Live Demo

Production deployment (Vercel): https://unmutte.vercel.app

Use this link to quickly verify UI changes, AI chat responses, and performance in a production-like environment. Note that environment variables for AI provider (Gemini/NVIDIA) must be set in the Vercel project for the chat to return real model output.

If you want the AI chat to work locally (calls /api/chat), run the Netlify function runtime:

npm run dev:netlify

Notes:

• This uses npx netlify-cli@17.34.0 under the hood. If you see errors, use Node 18–20 (e.g. nvm use 20).
• Create .env from .env.example and set GEMINI_API_KEY for local function dev. Never commit real secrets.

Build for production

npm run build

The static files are output to the dist/ directory (see vite.config.ts outDir).

Preview the production build

npm run preview

This serves the built site locally (default http://localhost:4173). Useful to validate the production bundle.

AI Chat providers

The AI chat uses a Netlify Function (/.netlify/functions/gemini-chat) that proxies requests to your chosen provider so API keys stay server-side.

Supported providers:

• Google Gemini (default)
• NVIDIA API (Nemotron, etc.) via OpenAI-compatible Chat Completions

Configure via environment variables (either in .env for local dev or in Netlify site settings):

• AI_PROVIDER — gemini (default) or nvidia
• If gemini:
  • GEMINI_API_KEY=<your_gemini_key>
  • Optional: GEMINI_MODEL=gemini-2.5-flash
• If nvidia:
  • NVIDIA_API_KEY=<your_nvidia_key>
  • Optional: NVIDIA_MODEL=nvidia/nemotron-4-9b-instruct

The frontend posts to /api/chat, which is redirected to the function in netlify.toml.

Deploying on Vercel

If you host on Vercel instead of Netlify, this project includes a Vercel Serverless Function at api/chat.js. Requests to /api/chat will automatically invoke the function on Vercel.

Set these environment variables in your Vercel Project (Production and Preview as needed):

• AI_PROVIDER = gemini or nvidia

• GEMINI_API_KEY (when using Gemini)

• GEMINI_MODEL (optional; the function auto-discovers supported models)

• NVIDIA_API_KEY (when using NVIDIA)

• NVIDIA_MODEL (optional; the function falls back to widely-available models if restricted)

• VITE_MIXPANEL_TOKEN (optional) — your Mixpanel project token. Prefix with VITE_ so the token is exposed to the client by Vite. Keep this token secret in repo; set it in Vercel/Netlify UI.

Build: npm run build (detected automatically) Output dir: dist (Vite default here)

Troubleshooting on Vercel:

• If you see "AI model not available" in the UI, it often means /api/chat returned 404/400. Ensure api/chat.js is deployed and your environment variables are configured. Check Vercel Function logs for details.
• Some NVIDIA models may be restricted per account/region. If you get 404/403 from NVIDIA, set NVIDIA_MODEL=meta/llama-3.1-8b-instruct or leave it empty to let the function try safe fallbacks.

Requirements

• Node.js 18+ recommended
• Modern browser

Troubleshooting

• If port 3000 is taken, change the server.port in vite.config.ts or run: npm run dev -- --port 4000.
• After dependency changes, delete node_modules and reinstall: rm -rf node_modules package-lock.json && npm install.