Merlin is built around a simple idea: instead of opening a separate chat tab for every task, you summon one assistant with a click while staying inside whatever you are reading or working on. The extension injects a sidebar on any web page, so it can read the article, the PDF, the YouTube transcript or the Gmail thread you currently have open and respond with that context already loaded. That makes summarization, rewriting, replying and research feel like part of the page rather than a separate tool.
Under the hood Merlin routes requests across a broad set of models — OpenAI's GPT family, Anthropic's Claude, Google's Gemini, DeepSeek, Mistral, Meta's Llama and others — so you can pick the model that suits a task without holding separate subscriptions for each. Beyond chat there is an image generator with multiple models and aspect ratios, a search mode that answers questions from live web results, document and video summarization, and a library of saved prompts. Newer additions like Projects (reusable knowledge bases) and Crafts (generated diagrams, mind maps and charts) push it toward light research and prototyping work.
Merlin is available as a Chrome extension, a web app, an iOS app and an Android app with synced history, custom prompts and custom bots across devices. The vendor highlights GDPR, ISO 27001 and SOC 2 alignment for data handling, and positions itself as accessible in 200+ countries where some other assistants are restricted. It is squarely a productivity assistant rather than a developer platform.