Reply to dozens of tweets per day in a chosen voice such as Sarcastic or Insightful
Generate personalized LinkedIn connection messages from a template
Tune the system prompt to match your own X persona
Route requests through OpenRouter instead of OpenAI directly
Needs a paid OpenAI or OpenRouter API key stored in Chrome local storage before any template button will produce a reply.
This repository holds a Chrome browser extension called ChatterBox that uses an AI model to write replies for you on X (formerly Twitter) and LinkedIn. The repository name on GitHub is xgrower-extension and the description credits GrowthHunt.ai, but inside the code and README it goes by ChatterBox. When you click the reply button on a tweet, the extension adds template buttons next to the reply box. You pick a template like Question, Funny, Agree, Sarcastic, or Insightful, and the extension calls the AI to generate a reply in that style, which you can then edit before posting. The extension supports OpenAI models such as GPT-5, GPT-4, and GPT-4o. You bring your own API key, which is stored locally in Chrome's storage rather than on any server. The settings popup also lets you change the system prompt that shapes the AI's voice, pick a model, and adjust advanced parameters like temperature, maximum tokens, presence penalty, and frequency penalty. The troubleshooting section mentions OpenRouter as well, which suggests the project may also route through that service. There are about ten built-in X templates, covering questions, jokes, agreement, sarcasm, insight, disagreement, congratulations, generic replies, and encouragement. For LinkedIn there are connection-message templates aimed at personalized outreach. The code is written in TypeScript and bundled with webpack. To install it, you clone the repository, run npm install and npm run build, then load the unpacked extension folder through Chrome's developer mode page. The project is MIT licensed and has 20 stars.
Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.