Cabin Chat

Voice-first multi-AI chat built for big screens — especially Tesla’s in-car Chromium browser with microphone support.

Talk or type. Replies from OpenAI, Claude, Gemini, or xAI Grok. Optional browser TTS or ElevenLabs.

Quick start

npm install
npm start

Open http://localhost:3000, pick a provider in Settings, save an API key (or set env vars), then tap Talk.

Environment variables

Set these for local (.env.local) or on Vercel (Project → Settings → Environment Variables):

Variable	Required	Purpose
OPENAI_API_KEY	One provider key	OpenAI / ChatGPT
ANTHROPIC_API_KEY	One provider key	Claude
GOOGLE_API_KEY	One provider key	Gemini
XAI_API_KEY	One provider key	Grok
PREFERRED_PROVIDER	Optional	openai, claude, gemini, or xai
TTS_PROVIDER	Optional	browser (default) or elevenlabs
ELEVENLABS_API_KEY	Optional	Premium TTS
ELEVENLABS_VOICE_ID	Optional	Default ElevenLabs voice
PORT	Optional	Local server port (default 3000)

On Vercel, prefer setting keys in the dashboard. The in-app Settings save updates keys for the current serverless instance only (filesystem is read-only).

Tesla tips

• Use HTTPS (Vercel production URL) so microphone permissions work
• Allow mic access when Chromium prompts
• Park before long chats; keep interactions short while driving
• Prefer Auto-speak + browser TTS for hands-free replies

Stack

• Vanilla HTML / CSS / JS
• Node.js + Express
• Web Speech API (recognition)
• Deployable on Vercel as a Node serverless function

Scripts

• npm start — production server
• npm run dev — nodemon