PaperLens ChatGPT Web Bridge

PaperLens ChatGPT Web Bridge connects the locally running PaperLens reader to a ChatGPT conversation that remains visible in the user's Chrome browser. When the user explicitly starts a PaperLens task, the extension can: • Find or open a visible chatgpt.com tab. • Insert the current document context, selected passage, question, or a structured translation and terminology prompt into the visible Chat composer. • Upload only images that the user explicitly attached in PaperLens. • Wait until ChatGPT has stopped generating and the visible answer body has remained stable before returning that answer to PaperLens. The extension supports PaperLens question answering, page and full-document translation, terminology extraction, selected-passage explanation, and image-aware reading tasks. Translation and terminology use strict structured prompts so PaperLens can map the visible ChatGPT response back to the correct page segments. Privacy and security: • The extension does not request or store ChatGPT passwords, cookies, API keys, or authentication tokens. • It does not use private ChatGPT APIs or hidden background conversations. • It operates only on the visible ChatGPT page after a user-initiated PaperLens request. • A short-lived local pairing token restricts the bridge to the PaperLens instance running on the same computer. • Prompt content and explicitly attached images are sent only through the user's visible ChatGPT conversation. • The extension does not sell user data or use it for advertising, credit decisions, or unrelated purposes. PaperLens Local is open source. The local service and setup instructions are available at: https://github.com/Peter-cuhk/PaperLens