Code-first browser automation, built and maintained by AI @ycombinator S22 company

WA, USA
Joined June 2022
Cloudflare launched a new model 𝚌𝚕𝚊𝚒𝚛𝚒𝚌 today that only does one job: extract values. We threw 86 real web pages at it and compared it to Haiku and Luna. The speed difference was real. Everything else favored the LLMs 🧵
1
2
5
209
Token count was the quiet killer. Same page was ~20.7k tokens for Clairic vs ~8.3k for Luna. Same price per token, so Clairic cost more.
1
47
We build this kind of extractions every day at Intuned — code-first scrapers and browser automations that an AI agent writes and maintains.
40
We're open-sourcing Selector Forge - an AI browser extension that generates semantic, resilient CSS & XPath selectors 🚀 🌐Creating reliable selectors is still one of the most painful parts of browser automation Free for 200 selectors/month and open source on github!
5
6
29
7,587
Selector Forge on Chrome: chromewebstore.google.com/de…... Selector Forge on Firefox: addons.mozilla.org/en-US/fir…... Selector Forge code: github.com/Intuned/selector-…
120
TikTok Shop's slider captcha vs. one line of code. `wait_for_captcha_solve()` - that is it! Check out the video📷
2
4
228
Intuned retweeted
Replying to @IntunedHQ
@IntunedHQ is live on HN today!
1
1
3
151
Intuned retweeted
Your scraper runs perfectly on localhost. You deploy it and start seeing CAPTCHAs you never saw before and you can't bypass them no matter what you do. The reason is you are being marked by A bot detection system. Here is how those systems actually work, layer by layer:
3
1
4
233
Intuned retweeted
The hardest part of browser automation isn’t building it. It’s maintaining it. Every time a website changes, your automation breaks. Debugging it manually is painful, time-consuming, and kills velocity. That’s why we built “Fix with AI” in @IntunedHQ . Because what you deploy and run in Intuned is actual code, we know exactly what ran, what failed, and why. So with one click, Intuned Agent analyzes the failure and proposes a precise fix - no guessing, no manual debugging. Watch how it works 👇
12
7
47
240,958
Your scraper runs perfectly on localhost. You deploy it and start seeing CAPTCHAs you never saw before and you can't bypass them no matter what you do. The reason is you are being marked by A bot detection system. Here is how those systems actually work, layer by layer:
3
1
4
233
Intuned retweeted
Last week Microsoft dropped Webwright — a terminal-native web agent framework using Playwright. Webwright just achieved SOTA results on Odysseys and Online-Mind2Web with a 100-step budget. We were excited to try it out… and honestly blown away by the results. Today we’re releasing the webwright-to-intuned skill! Deploy your Webwright crafts instantly and get a clean API you can call from anywhere. ✅No bot detection ✅No captchas ✅Intuned agent automatically maintains your scripts if they break ✅ Full observability & monitoring Turn one-off crafts into production-grade web agents in minutes. 🚀
4
9
1
141
27,949
Intuned retweeted
Scrapers don't fail loudly. They fail silently, in prod, and you find out way too late. self-healing projects on @IntunedHQ : detect → diagnose → fix → deploy. Autonomy is a dial you control. See how it works → intunedhq.com/docs/main/02-i…
1
4
3
194
Scrapers don't fail loudly. They fail silently, in prod, and you find out way too late. self-healing projects on @IntunedHQ : detect → diagnose → fix → deploy. Autonomy is a dial you control. See how it works → intunedhq.com/docs/main/02-i…
1
4
3
194
Intuned retweeted
Another example
1
1
3
338