Launched this week
Qursor
Point at any UI to send exact context to your AI
549 followers
Point at any UI to send exact context to your AI
549 followers
I kept wasting AI tokens describing UI changes to agents that edited the wrong element. So I built Qursor. Point at any element, copy structured context (selectors, classes, styles, fonts, colors), paste into your AI agent. No vague screenshots. No burned credits. - Inspect fonts, colors, spacing - Copy AI-ready element context - Extract components as HTML/CSS/JSX - Color picker and font detector - Download assets from any page
Free Options
Launch Team / Built With
Just use MСP Сhrome Dev Tools for your AI agent and it will be able to see the page, no need for any additional services that transfer the UI into the context
Qursor
@dren0r Totally, MCP DevTools is great when your agent already has full page access wired up. Qursor is more for cases where you are just chatting with an AI coding agent or using different tools, and you want a fast, framework‑agnostic way to point at a specific element and hand over clean selectors, styles or HTML without wiring anything.
@theomkarbirje Claude has chrome extension that can view the same page as you https://chromewebstore.google.com/detail/claude/fcoeoabgfenejglbffodgkkbkcdhcgfn. You can chat with page in side panel if you prefer use chat instead agents
Hey, that's smart, great idea. I use Tidewave, and that includes a pointer as well, but it only works with certain frameworks, so your extension might be what I need for other projects, I'll check it out!
Qursor
@jaime_iniesta Tidewave looks great when its pointer works with your stack, and Qursor is aiming to be the “works everywhere” version so you can point at any UI in Chrome and still grab clean selectors and styles for your agent, even on projects outside those supported frameworks.
This is a real agent failure mode. More context is useful, but exact target context is what stops the agent from confidently editing the wrong thing.
Curious if you see Qursor staying as copy-to-agent context, or moving toward a guardrail where the agent can propose the patch and leave a receipt of exactly what changed.
The token-waste problem is real. How does Qursor handle stateful elements - like a dropdown that's only visible when open, or a hover state? Those are usually impossible to inspect without freezing the UI somehow.
I vibe-code most of my app with Claude Code and the "no, the other blue button" loop is painfully real. Copying selectors + styles instead of a vague screenshot is the right fix. Any plans for a Safari version, or is Chrome the long-term home?
This is a great idea especially for people who want to create websites. Personally I've been using claude code, Claude design specifically. I was wondering what separates this from the mark up feature in Claude design where you just circle the part you want to change and send a prompt through to Claude.
I can definitely see this as being very useful for non-website development specific AIs
I spend way too much time copy-pasting error messages and UI screenshots into Claude just to explain what I'm looking at. If this actually captures the right context automatically that's a huge time saver. How does it handle dynamic content like dropdowns or modals that disappear when you click away?