Prompt Porter

Capture prompts and file attachments from one AI platform and replay them on any other - in one click.

Chrome extension link: Prompt Porter

---

What it does

When you're comparing AI models, you constantly retype the same prompt and re-attach the same files. Prompt Porter eliminates that. It intercepts your prompt text and file attachments, stores them as a bundle, and lets you inject that bundle into any supported platform with one click - or open multiple platforms simultaneously for side-by-side comparison.

Key features

• Capture - snapshot your prompt text and attached files from any supported LLM page
• Inject - replay a bundle into a different platform, populating the text field and re-attaching files automatically
• Compare mode - open 2-4 platforms in parallel tabs and inject the same bundle into all of them at once, with optional auto-send
• Keyboard shortcuts - Ctrl+Shift+S / ⌘+Shift+S to capture, Ctrl+Shift+V / ⌘+Shift+V to inject the latest bundle
• Local storage only - bundles are stored in your browser (IndexedDB + chrome.storage.local); nothing is sent to any server
• Auto-expiry - bundles older than 7 days are pruned automatically; max 20 bundles kept at a time

---

Supported platforms

Platform	Capture	Inject text	Inject files
ChatGPT (chatgpt.com)	✅	✅	✅
Claude (claude.ai)	✅	✅	✅
Gemini (gemini.google.com)	✅	✅	✅ ¹
Grok (grok.com)	✅	✅	✅

¹ Gemini requires an extra click due to browser security restrictions on file pickers without direct user interaction. After clicking Inject in the popup, a green Inject Bundle button appears on the Gemini page - click it to complete the injection.

---

Installation

From the Chrome Web Store (recommended)

Search for Prompt Porter on the Chrome Web Store and click Add to Chrome.

Manual / developer install

1. Clone or download this repository.
2. Open Chrome and go to chrome://extensions.
3. Enable Developer mode (toggle in the top-right corner).
4. Click Load unpacked and select the repository folder.
5. The Prompt Porter icon will appear in your toolbar.

---

How to use

Capturing a bundle

1. Open any supported AI platform and type your prompt. Attach files if needed.
2. Click the Prompt Porter toolbar icon.
3. Optionally enter a label for the bundle.
4. Click Capture (or press Ctrl+Shift+S / ⌘+Shift+S). The bundle is saved locally. A badge briefly confirms the capture.

Injecting into a single platform

1. Navigate to your target AI platform.
2. Open the Prompt Porter popup and select a bundle from the list.
3. Click Inject Here. The prompt text and files are automatically inserted into the platform's input area.

Comparing across platforms

1. Select a bundle from the list.
2. In the Compare section, check the platforms you want to open.
3. Optionally enable Auto-send to submit the prompt automatically after injection.
4. Click Compare. Prompt Porter opens a new tab for each selected platform and injects the bundle into all of them.

---

File structure

prompt-porter/
├── manifest.json       # Extension manifest (MV3)
├── background.js       # Service worker: storage, routing, keyboard shortcuts
├── content.js          # Content script (isolated world): capture, relay
├── injector.js         # Main-world script: React/Angular-aware text & file injection
├── adapters.js         # Per-platform DOM selectors (used during inject)
├── popup.html          # Extension popup UI
├── popup.js            # Popup logic: bundle list, capture, inject, compare
└── icons/
    ├── icon16.png
    ├── icon48.png
    └── icon128.png

---

Permissions

Prompt Porter requests only the permissions it needs:

Permission	Why
storage	Saves bundle metadata to chrome.storage.local
tabs	Opens new tabs for compare mode; queries the active tab
activeTab	Sends messages to the current tab's content script
Host permissions for the 4 supported sites	Runs the content script and injects bundles

No network requests are made. No data leaves your browser.

---

Privacy

All captured data - prompt text and file contents - is stored locally in your browser only. See PRIVACY.md for the full privacy policy.

---

Contributing

Pull requests are welcome. For significant changes, please open an issue first to discuss what you'd like to change.

When adding a new platform:

1. Add its hostname to manifest.json (content_scripts matches and host_permissions).
2. Add an adapter entry in adapters.js with getTextarea, getFileInput, getFileDropZone, injectText, and injectFiles.
3. Add a matching entry in the ADAPTERS object in content.js with getTextarea.
4. Add platform-specific injection logic in injector.js if needed (React fiber walking, shadow DOM, etc.).

---

License

MIT.

---

Built by Akash Tamate