A Chrome extension that lets you talk to Google's Gemini AI with voice and share your screen for visual context.
- Voice Chat: Talk to AI with your voice and get audio responses
- Screen Sharing: Share your current tab or screen with the AI
- Live Conversations: Voice + screen sharing together for smart assistance
- Go to Google AI Studio
- Sign in with your Google account
- Click "Create API Key"
- Copy the API key
- Load the extension in Chrome
- Open the side panel
- Paste your Gemini API key into the setup field
- Click
Save
- Open Chrome and go to
chrome://extensions - Turn on "Developer mode" (top right toggle)
- Click "Load unpacked"
- Select the
Extensionfolder - The extension will appear with a blue icon
This project does not use a bundler. The build process is just packaging the contents of the Extension folder into distributable Chrome extension files.
- Google Chrome installed at
/Applications/Google Chrome.app - The extension source in the
Extensionfolder - The signing key at
build/GeminiChrome.pemif you want to keep the same extension ID for the.crx
build/GeminiChrome.zipfor distribution or manual inspectionbuild/GeminiChrome.crxfor a signed packaged extension
From the repository root:
cd /Users/pablo/Projects/TheSeerCreate the ZIP from the extension source:
cd Extension
zip -qrFS ../build/GeminiChrome.zip . -x '*.DS_Store'Create the signed CRX with the existing key:
"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome" \
--pack-extension=/Users/pablo/Projects/TheSeer/Extension \
--pack-extension-key=/Users/pablo/Projects/TheSeer/build/GeminiChrome.pem
mv -f /Users/pablo/Projects/TheSeer/Extension.crx /Users/pablo/Projects/TheSeer/build/GeminiChrome.crx- Update the files inside
Extension/. - Rebuild the ZIP so the latest source is packaged.
- Run Chrome's
--pack-extensioncommand to generate a signed.crx. - Move the generated
Extension.crxintobuild/GeminiChrome.crx. - Load
Extension/unpacked in Chrome for development, or use the files inbuild/for packaged distribution.
- If you omit
--pack-extension-key, Chrome can still create a.crx, but the extension ID will change. - For local development,
Load unpackedfromExtension/is usually the fastest workflow.
- Click the extension icon to open the side panel
- Click the blue "Talk" button
- Allow microphone access when prompted
- Start speaking - the AI will respond with voice
- Click the red "Stop" button to end
- During voice chat, your screen is automatically captured every 3 seconds
- The AI can see what's on your screen and help accordingly
- Grant screen capture permission when prompted
"Please enter your API key" - Open the side panel and save a valid Gemini API key
Microphone not working - Allow microphone permission in Chrome settings
Screen capture fails - Allow screen recording permission when Chrome asks
Connection errors - Check your internet and verify your API key is valid
The extension can use Google Cloud Firestore to persist Gemini Live session resumption tokens. When enabled, the conversation context survives closing the side panel and reconnects (e.g. after the 2-minute context-window refresh). If Firestore is not configured, the extension still works; the token is kept only in memory and is lost when the panel closes.
- Where: Side panel → Settings → Firestore (optional) (Project ID and API key).
- What it stores: A single document (
gemini_sessions/current_session) with aresumptionTokenfield holding the Live API resumption handle. - Proof of use: See GOOGLE_CLOUD_FIRESTORE.md in the repo root for a full description of how Firestore is used, the REST API calls, and the code paths that read and write the token.
- Get API key: Google AI Studio
- Chrome Extensions:
chrome://extensions - Check browser console (F12) for error messages