A Chrome extension that generates contextual, human-sounding replies for X (Twitter) posts using AI.
- AI-Powered Replies: Generate 5 unique reply suggestions for any tweet
- Multiple Tones: Choose from 7 different tones (Sarcastic, Witty, Casual, Cool, Roasting, Encouraging, Professional)
- Adjustable Length: Short (1 sentence), Medium (1-2 sentences), or Long (2-3 sentences)
- Model Selection: Choose between text-only and vision-capable AI models
- Image Understanding: Vision models can see and understand images in tweets (memes, screenshots, infographics)
- One-Click Insert: Insert replies directly into the reply box
- Keyboard Shortcut: Quick access with
Alt+R - Dark/Light Mode: Automatically matches X's theme
- Context-Aware: Extracts thread context, quoted tweets, and media for better replies
- Character Count: Shows reply length with 280 character limit indicator
- Clone or download this repository
- Open Chrome and navigate to
chrome://extensions/ - Enable "Developer mode" (toggle in top right)
- Click "Load unpacked"
- Select the
reply-xfolder
- Get an API key from OpenRouter
- Click the Reply X extension icon or go to extension settings
- Enter your OpenRouter API key and save
- (Optional) Select your preferred AI model
- Visit X.com
- Find a tweet you want to reply to
- Click the Reply X button next to the share button
- Select your preferred tone and length
- Click Insert to add reply to the text box, or Copy to clipboard
Press Alt+R to open Reply X for the tweet in the center of your screen. Press again to close.
Choose your preferred model in Settings:
| Category | Model | Cost/Request | Features |
|---|---|---|---|
| Text Only | Llama 3.1 8B | ~$0.00005 | Fast, cheapest option |
| Text Only | Mistral 7B | ~$0.00005 | Fast, good quality |
| Text Only | Gemma 2 9B | ~$0.00006 | Google's efficient model |
| Vision | GPT-5 Nano | ~$0.00015 | Can understand images |
| Vision | Llama 3.2 11B Vision | ~$0.00006 | Open source with vision |
| Vision | Gemini 2.0 Flash Lite | ~$0.00008 | Fast vision model |
Vision models can understand images in tweets, making them ideal for:
- Memes and screenshots
- Infographics and charts
- Photos with context
| Tone | Description |
|---|---|
| Sarcastic | Light sarcasm and dry humor |
| Witty | Clever, quick replies with wordplay |
| Casual | Natural conversational tone |
| Cool | Internet-native, unbothered energy |
| Roasting | Playful burns and teasing |
| Encouraging | Supportive and positive |
| Professional | Clear and thoughtful |
Set your default tone and reply length in the settings page. These will be pre-selected when you open the popup.
- Your API key is stored locally in Chrome storage
- Tweet data (text, images) is sent to OpenRouter for AI processing
- No data is stored on external servers by this extension
- Images are only sent when using vision-capable models
- Chrome Extension Manifest V3
- OpenRouter API (Multiple model support)
- Vanilla JavaScript
- CSS with dark mode support
MIT