Gemini Context Actions (Chrome Extension)

A Chrome extension that adds Leo-style context-menu actions for any selected text, powered by Google Gemini.

• ❓ Ask a question → type a question about the selected text and get an answer from the model (uses the selection as context).
• 🌍 Translate → English, French, German, Italian, Spanish (shown first in the menu).
• 🖍️ Highlight → Yellow / Orange / Red / Green / Cyan / Purple (inline highlights with a tiny “×” bubble to remove).
• ⚡ Quick actions → Summarize, Explain.
• ✍️ Rewrite → Paraphrase, Improve, Change tone → Academic / Professional / Persuasive / Casual / Funny.
• 📏 Change length → Shorten, Expand
• 🎯 Create → Tagline, Social media → Short post / Long post.

The result appears in a floating on-page panel with Copy, Replace selection, Settings, and Close controls.

---

📸 Screenshots

❓ Ask a question:

🌍 Translate:

🖍️ Highlight:

---

📦 Project structure

gemini-context-actions/
├─ manifest.json
├─ background.js
├─ contentScript.js
├─ contentStyles.css
├─ options.html
├─ options.js
└─ icons/
   ├─ icon16.png
   ├─ icon24.png
   ├─ icon32.png
   ├─ icon48.png
   ├─ icon128.png
   └─ icon.svg

---

🚀 Install (from source)

1. Clone or download this repository.
2. Open chrome://extensions.
3. Enable Developer mode (top-right).
4. Click Load unpacked and select the project folder.

---

🔑 Add your Gemini API key

1. In chrome://extensions, find Gemini Context Actions → Details → Extension options (or right-click the toolbar icon → Options).
2. Paste your Google Gemini API key (from Google AI Studio) and click Save.
3. Click Test to verify the key—if it succeeds, you’re good to go.

Your key is stored in chrome.storage.sync and used only by the background service worker to call the Gemini API over HTTPS.

---

🖱️ Use it

1. Select text on any webpage.
2. Right-click to open the context menu.
3. Choose Translate, Quick actions, Rewrite, Change length, or Create.
4. A floating panel shows the result. Use Copy to clipboard or Replace to overwrite the current selection.

---

🔁 Changing the model (depending on the API you have)

This extension talks to Google AI Studio’s Generative Language API (REST v1) by default. If your key has access to newer or different models, update the constant in background.js.

A) You have an AI Studio (Generative Language API) key ✅ (recommended)

• Endpoint (already set in the code): https://generativelanguage.googleapis.com/v1
• Open background.js and set:

// background.js
const API_BASE = "https://generativelanguage.googleapis.com/v1";
const MODEL = "gemini-2.5-flash-lite"; // default; change as you like

Common model IDs (as of today):

• gemini-2.5-flash-lite — fastest & cheapest, great for rewrite/translate/summarize.
• gemini-2.5-flash — balanced quality/speed.
• gemini-2.5-pro — highest quality; slower & pricier.

Notes:

• Avoid old -latest aliases (e.g., gemini-1.5-flash-latest) and avoid v1beta unless you know you need it.
• If you switch the default model, you don’t need to change anything else in the extension.

Also update the Test button in options.js if you want it to ping your chosen model:

// options.js (in test() function)
const url = "https://generativelanguage.googleapis.com/v1/models/gemini-2.5-flash-lite:generateContent?key=" + encodeURIComponent(k);

B) You only have a Vertex AI key (Google Cloud) ⚠️

Vertex AI uses different auth (OAuth/Service Account) and different endpoints. Browsers can’t safely sign these requests directly (no secret storage / CORS). If you must use Vertex AI, set up a minimal proxy (Cloud Run/Functions) that:

• accepts { model, contents } from the extension,
• calls Vertex AI server-to-server with your service account,
• returns the generated text.

Then, in background.js, point API_BASE to your proxy instead of Google:

// background.js (if using your proxy)
const API_BASE = "https://your-proxy.example.com"; // your HTTPS proxy
const MODEL = "google/gemini-2.5-flash"; // Vertex model name (your proxy maps it)

If you don’t want to run a proxy, use an AI Studio key instead. That’s the path this extension supports out of the box.

---

🛡️ Permissions

• contextMenus — add the right-click menu.
• activeTab, scripting — communicate with the current page (to render results / replace text).
• storage — store your API key.
• host_permissions for https://generativelanguage.googleapis.com/* — talk to the Gemini API.

---

🔒 Privacy & security

• Your API key is saved via chrome.storage.sync and never injected into the web page.
• Requests are performed in the background service worker.
• No analytics or third-party calls besides the Gemini API.

---

🛠️ Troubleshooting

• 404 NOT_FOUND with -latest models Use the v1 endpoint and a concrete model ID (e.g., gemini-2.5-flash-lite, gemini-2.5-flash, gemini-2.5-pro).
• “API key required” in the floating panel Open Options and save your key.
• Still stuck? Check chrome://extensions → Service Worker console for the extension.

---

🧑‍💻 Development

• Edit code → chrome://extensions → Reload to pick up changes.
• The floating panel’s look & feel is in contentStyles.css.

---

📄 License

Copyright © 2025 Vittorio Pio Remigio Cozzoli. This project is licensed under the MIT License. You are free to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the software under the terms of the MIT license.

See the full text in LICENSE file included in this repository.

---