You find a great article. You want to keep the useful part, not the newsletter popups, the cookie banner, and the fourteen "related posts" cards. So you copy-paste it into a note, spend two minutes stripping out garbage formatting, manually type the source URL, and give up on the author's name. Multiply that by every page worth saving and you stop saving pages at all. Your "read later" pile becomes a browser window with ninety tabs you will never reopen.
The official Obsidian Web Clipper exists to kill that friction. It's a free browser extension for Chrome, Firefox, Safari, and Edge that reads the meaningful content off a page and drops it into your vault as Markdown, locally, with no account and no telemetry. As of mid-2026 it's the default way to get web content into Obsidian, and the interesting part isn't the clipping itself. It's the template engine underneath it.
Templates that match the site you're on
A template defines the shape of the note a clip produces: its title, which folder it lands in, what properties it gets, and how the body is formatted. You can have as many templates as you want, and this is where it stops being a glorified copy button.
Each template can carry triggers that auto-select it based on the current page URL or the page's schema.org data. Define a rule per line, and the first template whose trigger matches wins. Clip a recipe site and you get a recipe template; clip an arXiv page and you get a research template with the abstract and authors already broken out; clip anything else and a sensible default catches it. You set this up once and then never think about which template to pick again.
Variables and filters do the formatting for you
Templates are populated by variables, placeholders that get replaced with real page data at clip time. They come from several sources: preset variables like {{title}}, {{url}}, and {{date}}; meta tag variables pulled from the page's HTML <meta> elements; selector variables you extract with a CSS selector; and schema.org variables lifted from structured data the site already publishes.
Filters then transform those values, using a pipe syntax borrowed straight from templating languages you've probably seen: {{variable|filter}}, chainable with more pipes. There are over fifty of them, so you can slugify a title, format a date, trim whitespace, or split author lists without touching the note after it lands.
A template's properties section is just YAML frontmatter with variables in it. Something like this fills itself in on every clip:
---
title: "{{title}}"
source: "{{url}}"
author: "{{author}}"
published: { { date|date:"YYYY-MM-DD" } }
tags: [clippings]
---Because those are real Obsidian properties, everything you clip is immediately queryable. A Dataview or Bases view over source and published turns your clippings folder into a searchable reading log instead of a junk drawer.
Highlighter, Reader, and Interpreter
Three companion features round out the extension. The Highlighter lets you mark passages on a live page and save only those; return to the page later and your highlights are still there, which makes it a genuine research tool and not just an archiver. Reader strips a page down to a clean, distraction-free view before you clip, so the noise is gone from the start.
Interpreter is the one that feels new. It runs natural-language prompts against the page at clip time, so you can auto-generate a summary, pull out the key claims, or extract a structured list into a variable, and pipe that straight into your template. Clip a dense whitepaper and have the note arrive with a three-bullet TL;DR already at the top. Note that Interpreter sends page content to whatever model provider you configure, so it's the one piece here that isn't fully local. The core clipping stays on your machine.
Where to start
Install the extension, then resist the urge to build the perfect template on day one. Start with the default, clip a dozen pages, and notice what you keep fixing by hand. That annoyance is your spec: add a property here, a trigger there, a filter to format the date, until clipping is a single click that produces a note you'd actually reread. The official documentation covers the full variable and filter reference, and the community template collection is worth a look before you write your own from scratch.