Build a scraping pipeline without writing code
A recurring scraping pipeline by hand is a browser script, hand-written selectors, a proxy pool, and a babysat cron job. SmartBrowse builds it from clicks.
Say you need the same data every morning: the new listings on a page that loads them as you scroll, across a dozen paginated screens. By hand, that is a script to drive the browser, a set of selectors to pull each field, a proxy so the site does not block you, a schema to validate the output, and a cron job with retries and an alert for when it breaks at 3am. Most of that work has nothing to do with the data you wanted. It is the machinery around it.
SmartBrowse is the machinery. You open a recipe, click the first two items you want, and it infers the repeating pattern. Save it, and the recipe runs on demand or on a schedule. There are no selectors to write and no script to keep alive through the next redesign.
How it works
You demonstrate the action instead of scripting it. Click an example or two of
what you want.
SmartBrowse watches the selection and works out the pattern that covers the rest of
the page. The thing you would have encoded as div:nth-child(3) > span you instead
just point at. Why described fields outlast hand-written selectors is its own
story, told in structured extraction vs CSS
selectors; the same logic is what
lets a couple of clicks stand in for a brittle query here.
Every run drives a real browser. Pages that need JavaScript, pagination, or a click to reveal content behave the way they do for a person, because a real browser is what runs them. Moving through the site is part of the recipe, not a separate scripting problem bolted on the side. The fetch tier underneath is the same one that escalates only as far as a page forces it.
Replay tolerates drift. On each run, replay tries your exact selection first and falls back to the page's structure when the layout has shifted. A run only bills when it completes, so a recipe that breaks on a redesign costs nothing until you re-record it.
Schedules and webhooks close the loop. Put a recipe on a recurring schedule and it fires on its own; a completed run can POST its result to your webhook. Schedules are available on Hobby and higher plans; on the free tier you run recipes whenever you need them.
What a run costs
A SmartBrowse run is 2 credits per page extracted, charged on completion only. Building and editing a recipe is free, so you can get the selection right before you ever spend a credit running it. It draws on the same balance as your scrapes and extractions, which means a recipe is one more line in one wallet, not a separate bill for browser time.
Who it's for
Someone who needs the data on a schedule but does not want to own a scraper: an analyst, an ops lead, a founder pulling a competitor's catalog every week. And the engineer who would rather not have a brittle click-through script for one paginated dashboard sitting in the on-call rotation, breaking quietly the week they are out.
When SmartBrowse is the wrong tool
If the data sits on a single static page, a one-call SmartScraper is simpler and cheaper: describe the fields, get JSON back, no clicking involved. SmartBrowse earns its 2 credits a page when you have to move through the site to reach the data: paginate, expand a section, filter, click through. It is also built around one authoring sitting, with a 15-minute cap on a live editing session, so it fits recipes you can demonstrate in a single pass rather than an afternoon of exploratory clicking.
Where this goes next
The machinery around the data was always the expensive part. Pushing it down to a couple of clicks is the whole feature, and it is the part you stop maintaining. Open a new recipe, click through the page once, and decide whether it runs when you ask or every morning while you sleep. Authoring is free and a failed run costs nothing, so the first one is yours to try at no risk.
Frequently asked questions
What is SmartBrowse?
SmartBrowse is a visual recipe builder that drives a real browser. You click the first couple of items you want on a page, it infers the repeating pattern, and you save that as a recipe you can run on demand or on a recurring schedule. No selectors, no script.
Do I need to write code or selectors to use SmartBrowse?
No. You demonstrate what you want by clicking an example or two, and SmartBrowse works out the pattern that covers the rest of the page. There is nothing to script and no CSS or XPath selector to write or maintain when the page changes.
How much does a SmartBrowse run cost?
Each run is 2 credits per page extracted, charged on completion only, so a run that fails costs nothing. Building and editing a recipe is free. The same credit wallet that covers your scrapes and extractions covers SmartBrowse runs too.
Can SmartBrowse run on a schedule?
Yes. A saved recipe can run on a recurring schedule, and a completed run can POST its result to your webhook. Schedules are available on Hobby and higher plans; on the free tier you run recipes on demand whenever you need them.
What happens when the target site changes its layout?
Replay tries your exact selection first and falls back to the page's structure when the layout shifts. A run only bills when it completes, so a recipe that breaks on a redesign costs nothing until you re-record it. You are not charged for runs that return no usable data.
Does SmartBrowse use a real browser?
Yes. Each run drives a fresh real browser, so pages that need JavaScript, pagination, or a click to reveal their content behave the way they do for a person. That is what lets a recipe move through a site instead of reading one static response.