Skimlane

Skimlane is a Chrome extension that processes web pages through AI and transforms them into structured summaries as you browse. Instead of reading every article, blog post, documentation page, or research paper in full, you define what information you want extracted and Skimlane pulls it out automatically. The output appears in a side panel that stays open while you navigate between tabs, giving you instant access to processed content without switching contexts, copying text into external tools, or losing your place. For anyone who spends hours reading online as part of their job, this changes how quickly you can move through information-heavy workflows.

The core mechanic is what the tool calls reading recipes. Each recipe is a template that tells the AI how to process a page. You pick which sections to extract, set the level of detail you want in the output, and define specific prompts for different content types. The extension ships with several premade recipes to get you started. Worth Reading judges whether an article deserves your full attention or can be safely skipped. PM Brief extracts key takeaways formatted for product managers who need to synthesize competitor updates or feature announcements. Signal Cleaner filters noise from dense pages, pulling out the core argument while ignoring filler paragraphs. You can use these out of the box, modify them to fit your needs, or build completely custom recipes from scratch. Recipes can be exported and imported, so teams can share processing templates across installations.

What makes Skimlane notable compared to similar tools is its flexibility around AI models. You're not locked into a single provider or forced to pay for cloud processing. The extension supports Chrome's built-in Gemini Nano for entirely on-device processing that never leaves your machine. It connects to local Ollama instances if you're running open source models on your own hardware. Groq's free tier provides cloud processing without subscription costs if you want faster results than local inference allows. And you can configure any OpenAI-compatible endpoint, including paid providers like OpenAI itself or Claude, if you want premium model quality. This architecture means you can process sensitive or confidential content entirely offline, use free tiers to avoid subscription fees during experimentation, or connect to high-end models when output quality matters more than cost.

Privacy sits at the center of the design philosophy. Skimlane describes itself as local-first by default, which isn't just marketing language. Your browsing history, processed summaries, recipe configurations, and saved outputs all stay in Chrome's local storage. There's no backend server collecting your data, no analytics tracking which pages you visit, no cloud database storing your processed content, and no account required to start using the extension. If you want to delete everything Skimlane knows about you, you uninstall the extension and your data disappears with it. For anyone who processes confidential documents, competitive intelligence, client research, or personal content they'd rather keep private, this architecture matters more than any feature list.

The practical workflow looks like this. You land on a page and your default recipe runs automatically, or you click the extension icon to trigger processing manually if you've disabled automatic mode. Results appear in the side panel as structured text formatted according to your recipe's instructions. From there you can save individual summaries as Markdown files for your notes, ask follow-up questions about the content using the AI model, or access your browsing history to revisit previously processed pages and compare summaries across sources. The side panel persists across tabs, so you can skim multiple articles in sequence and maintain context without constantly reopening windows or losing your place in the original documents.

Target users include knowledge workers drowning in research, product managers who need to synthesize competitor updates and industry news quickly, analysts processing large volumes of reports and filings, journalists monitoring multiple sources for developments, students working through reading lists, and anyone whose job involves extracting signal from noise across dozens of web pages daily. If you spend hours reading online and wish you could delegate the initial filtering to something smarter than browser find or manual highlighting, Skimlane fits that gap.

The extension is completely free with no sign-up required and no premium tier waiting to gate features. The combination of zero cost, local-first architecture, and model flexibility makes it unusually accessible compared to similar tools that charge monthly fees or require accounts. You can start with on-device processing using Gemini Nano, graduate to Groq's free cloud tier for faster results on longer documents, and eventually connect paid endpoints if your usage scales to the point where speed and quality justify the expense. The tool grows with your needs without forcing you into a subscription from day one.