browse.sh — an open catalog of browser automation skills for any website. Find reusable SKILL.md recipes that teach AI agents to complete tasks online, and install them with the browse CLI.
Over the last year, we've watched AI agents get remarkably good at using browsers. But we've also noticed something strange: every time an agent visits a website, it starts from zero.
It re-explores the interface, re-discovers buttons, re-learns navigation paths, and re-finds the same workflows it already completed yesterday.
Humans don't work that way.
Once you learn how to search Zillow listings, review a GitHub PR, or book a campsite on Recreation.gov, you don't relearn the entire website every time you come back.
Agents shouldn't have to either.
That's why we built Browse.sh, an open catalog of browser skills that agents can install and reuse across the web. Instead of exploring a website from scratch, agents load the relevant skill and execute against a known workflow.
The result is faster execution, lower token costs, more reliable outcomes, and better multi-site workflows. Today, the catalog includes 250+ skills across real websites and applications, including partner skills like submitting reimbursements on Ramp, creating projects on Lovable, extracting document data on Reducto, and many more.
And when a skill doesn't exist yet, Browse.sh can create one.
Behind the scenes, Browse.sh is powered by Autobrowse, our system that runs tasks in real browsers, analyzes traces, DOM changes, network activity, screenshots, and failures, then continuously improves the workflow until it converges on a durable strategy.
Over time, one successful browser run becomes a reusable skill that anyone can install.
Browse.sh is open source, free to use, and available today.
We'd love your feedback:
- Which websites do your agents struggle with most today? - What skills should we add next? - What workflows are you automating with AI agents?
We'll be around all day answering questions. Thanks for checking us out 🤞
This makes a lot of sense. Agents relearning the same website flow every time feels like such a waste. Curious how you handle it when a site changes its UI slightly. Does the skill update itself or need manual fixing?
I like the idea of agents not starting from zero every time. Re learning the same website flow feels like wasted time and tokens. How do you handle skills when a website changes its layout or button names?
@shrey150 The idea that agents should have muscle memory like humans do is so obviously right that it’s wild nobody shipped this sooner. Watching an agent re-explore the same GitHub PR workflow for the tenth time feels like watching someone forget how to use a doorknob... and now there’s finally a fix. Yaay!
The "agents start from zero every time" framing is exactly right. I drive a browser agent every day for marketing routines and the recurring tax is re-finding the same comment box, the same upvote button, the same auth modal. Reusable SKILL.md per domain is the unlock - and an open catalog turns it into a network effect. Following.
fire proposition. played around with browser automation to scrape scholarships off the web; wonder if something like this could help me apply to them automatically asw ...
This looks very interesting, definitely going to try a few of these skills this week.
Memory seems useful when an agent is actively working, but the harder problem feels like deciding what deserves to be remembered in the first place.
Have you found the biggest gains come from long-term memory across sessions, or from reducing context loss within a single workflow?
agents have no muscle memory, and that's a real cost in tokens, in time, in reliability.
the open catalog model is interesting. curious how skill quality is managed as it scales. is there a review layer or does it mostly rely on community signal to surface what actually works? good luck with the launch!
How much of an issue have you found captchas to be? Are websites trying to restrict agents or are they open to agentic browsing?
"Muscle memory for agents" nails the problem — re-deriving the same web flow on every run is where agent automations get slow and brittle. Congrats on the launch. Does the recorded memory adapt when a site's DOM shifts, or does it need a re-teach?
This makes a lot of sense. Teaching agents the same website every day feels a bit like deleting your browser history before every session 😄 Congrats on the launch!
The interesting part for me is not just reusing browser steps, it is making them reviewable.
For agent workflows I’d want to see how a skill handles the boring cases: site layout changes, failed clicks, partial completion, and logs that explain why the agent chose a path. That is usually where browser agents either become useful or become hard to trust.
The 'muscle memory' framing is apt: it's really a domain-specific replay buffer for web interactions. We've been building AI agents that automate customer workflows and session state management across sites is genuinely hard. How does Browse.sh handle sites that frequently change DOM structure? Does the agent re-learn from scratch or do partial cache invalidation?
Congrats the on the launch! We've been using Browserbase to benchmark https://github.com/pixiebrix/agent-browser-shield and prototyping how to enforce/apply the browse.sh content at the harness layer instead of context layer! Love it!
Anyone made a ryanair skill yet? I guess we need GPT 6 or 7 before its possible to navigate that website with AI 😬
There’s some level of stochasticity in the “perfect recipe” and resulting SKILL.md output for a given task. If multiple users each have a distinct generated skill to perform the same task uploaded to Browse.sh, how does the platform resolve these redundancies?
About Browse.sh on Product Hunt
“Give your agents muscle memory for automating the web”
Browse.sh launched on Product Hunt on June 8th, 2026 and earned 305 upvotes and 36 comments, earning #2 Product of the Day. browse.sh — an open catalog of browser automation skills for any website. Find reusable SKILL.md recipes that teach AI agents to complete tasks online, and install them with the browse CLI.
Browse.sh was featured in API (98.2k followers), Developer Tools (513.7k followers) and Artificial Intelligence (470.4k followers) on Product Hunt. Together, these topics include over 179.9k products, making this a competitive space to launch in.
Who hunted Browse.sh?
Browse.sh was hunted by fmerian. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how Browse.sh stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
fmerian
Hey Product Hunt 👋
I'm Shrey,
Over the last year, we've watched AI agents get remarkably good at using browsers. But we've also noticed something strange: every time an agent visits a website, it starts from zero.
It re-explores the interface, re-discovers buttons, re-learns navigation paths, and re-finds the same workflows it already completed yesterday.
Humans don't work that way.
Once you learn how to search Zillow listings, review a GitHub PR, or book a campsite on Recreation.gov, you don't relearn the entire website every time you come back.
Agents shouldn't have to either.
That's why we built Browse.sh, an open catalog of browser skills that agents can install and reuse across the web. Instead of exploring a website from scratch, agents load the relevant skill and execute against a known workflow.
The result is faster execution, lower token costs, more reliable outcomes, and better multi-site workflows. Today, the catalog includes 250+ skills across real websites and applications, including partner skills like submitting reimbursements on Ramp, creating projects on Lovable, extracting document data on Reducto, and many more.
And when a skill doesn't exist yet, Browse.sh can create one.
Behind the scenes, Browse.sh is powered by Autobrowse, our system that runs tasks in real browsers, analyzes traces, DOM changes, network activity, screenshots, and failures, then continuously improves the workflow until it converges on a durable strategy.
Over time, one successful browser run becomes a reusable skill that anyone can install.
Browse.sh is open source, free to use, and available today.
We'd love your feedback:
- Which websites do your agents struggle with most today?
- What skills should we add next?
- What workflows are you automating with AI agents?
We'll be around all day answering questions. Thanks for checking us out 🤞