EasyRead

EasyRead is a Chrome extension that adds summaries, in-context definitions, bilingual translation and Mind Maps to the page you are reading. It supports your own cloud API key, a local Ollama or LM Studio model, or Chrome built-in AI, with settings kept in the browser and no account required.

EasyRead is a browser extension for desktop Chrome (and other Chromium browsers that accept Chrome extensions) that adds reading assistance directly on top of the page you already have open. Instead of copying text into a separate chat window, you turn on Block Picker and click a passage, then ask for a summary, a translation, a definition in context, or a Mind Map of how the ideas connect. The original layout stays intact, and results appear in a floating panel or a sidebar, whichever you prefer. The design choices aim at two kinds of readers. The first is anyone working through material in a second language: bilingual translation keeps the source and the translation side by side, and you can translate the whole article or only a selected block. The second is the reader who wants a structural view of a long piece; the Mind Map can be zoomed, panned, expanded and collapsed inside the extension and exported as a PNG. Summaries and definitions are written to sit next to the passage they explain, so the reading flow is not broken by tab switching. What separates EasyRead from hosted reading assistants is where the model runs. Three setups are offered. Cloud BYOK connects OpenAI, Anthropic or another provider using your own API key, with usage billed by that provider. A local model service connects to Ollama or LM Studio running on your own machine, so there are no cloud API fees. Chrome built-in AI uses the browser's on-device model on supported desktop devices and needs no key at all. Settings and reading history are stored in the browser rather than on EasyRead servers, and no account is required. The trade-offs follow from that architecture. Answer quality and speed track whichever provider or local model you supply, and cloud use means task content travels to that provider. Local models put load on your machine's memory and processor, and the built-in model is described as best suited to simpler reading tasks, with a slower first reply after startup or idle periods. EasyRead is therefore closest to a lightweight reading layer over your existing AI stack rather than a self-contained subscription product.