Chat Overleaf is an AI chat assistant browser extension for the Overleaf online LaTeX editor. The project is developed based on the Plasmo framework and runs as a Chrome plugin. It aims to embed large language model chat capabilities into the Overleaf editing interface, helping users get AI assistance directly while writing and editing LaTeX documents.
In terms of main capabilities, the plugin supports asking questions directly after selecting text in the editor; the selected content is automatically used as context and comes with file path or directory information, making it easier for the model to locate and reference. Users can invoke the sidebar and focus the input box via the Ctrl + L shortcut or the selection floating button. The plugin can automatically obtain the current file content, and also supports manually extracting the entire project content as AI context; files are displayed in a tree structure, supporting folder expansion/collapse, flexible file selection, select all/clear, and showing the estimated Token count of selected files. When editor content changes, extracted files sync in real time, and the file list is cached by project ID.
For interaction and model configuration, the plugin comes preconfigured with multiple mainstream AI models such as DeepSeek, Kimi, Qwen, and Gemini, and also supports adding custom AI providers and models. After entering a model ID, the provider's model list can be automatically retrieved, and commonly used models can be pinned to the top. The chat supports multimodal input via image upload, paste, and drag-and-drop, can display the AI reasoning process, supports @ quick file references, and the interface is a resizable, responsive sidebar. Chat history is automatically saved in real time, supports loading and managing multiple sessions, and allows creating new conversation branches from historical messages.
For LaTeX editing, AI replies can be rendered as search/replace diffs, supporting regex or normal mode, and users can apply changes back to the Overleaf editor or reject them with one click; before applying, it can automatically jump to and highlight the area to be replaced, view it inline in a floating overlay, and applied or rejected changes can be undone and re-highlighted as candidates.
For usage, installation from the Chrome Web Store is recommended; you can also download the zip from GitHub Releases, extract it, enable Developer mode at chrome://extensions/, and load the chrome-mv3-prod folder (this method is marked as suspended for updates). After installation, you need to add a model key first, then visit Overleaf and click the icon in the lower right corner to use it. In development mode, you can clone the repository, use pnpm install to install dependencies, pnpm dev to startthe development server, and load the build/chrome-mv3-dev folder on the extension management page.
Applicable scenarios include Q&A, content extraction, and applying modification suggestions when writing LaTeX documents such as papers and reports in Overleaf. In terms of limitations, the information shows that it is a Chrome extension and does not provide support information for other browsers; charging, pricing, and specific model performance are not yet provided.
