Browser extension for scraping
A person opens a page under their own login, selects what matters, and the data lands in a table in a single shape. The next page of the same kind is collected without marking anything up again.
- Buildfrom $1,900
- Timeline2–4 weeks
- Supportfrom $140/mo
Who this suits: Work where what is needed is on the screen and does not come out of it: an account behind a login, a report with no export.
What is in the first version
- markup by selecting
- paging a list
- spreadsheet and JSON
- run history
The figures above are the package this sits on: Browser extension. Anything past its edges the calculator adds before the work starts, not after.
How the path goes
- 01
Show it once what to take
The extension remembers a place in the structure of the page rather than a spot on the screen.
- 02
It collects under your open session
It reaches exactly what the person reaches, stores no passwords and opens no second session.
- 03
It lands where it gets used
One shape throughout, so a second run joins the first, and every row carries its source address and the run time.
What decides whether this works or annoys people
A fast pass gives itself away
Pace and pauses are set to ordinary reading speed: an export that runs all night costs less than a blocked account.
An empty cell is not zero
A missing field is a marked row and a stopped run, not a blank cell that becomes a zero in a report.
The browser is the boundary
One-off and moderate volumes are an extension, regular collection on a schedule is a server, and the second costs different money.
What to measure once it is live
- pages collected in one run
- rows that need fixing by hand
- fields not found on a page
- time from a page to a finished export
Why this comes out faster
The shape of this one is known: the states, the edge cases and the things that usually go wrong have been decided before. Nothing here is a template, and the saving is not in your half of the work. It goes into your process, your content and the systems this has to talk to, which is the part nobody can have solved in advance.
A likely stack for this
Picked against the task when we scope it, not decided in advance. This is the shelf it usually comes off.
- ExtensionTypeScriptReactChrome Manifest V3
- Parsing and storageNode.jsPostgreSQLS3
- Checks and operationsPlaywrightDockerGrafana
Price it yourself, right here
Five steps, and you can see the number without leaving a contact. The estimate accounts for the kind of work, what you already have and what it will need inside.
These bills do not come from us
- paid subscriptions to the sources
- server and storage, if collection goes to a database
- AI model calls for unstructured descriptions
- a Chrome Web Store developer account
Asked before the first call
- How is this different from a parser on a server?
- The collection happens in your browser, under your session and on your press, so it reaches what you reach. Regular collection with nobody present is a different system with its own server.
- Is this legal?
- Reading a public page generally is. An account behind a login has terms the person accepts on registering, and personal data we do not collect at all. We read the terms before the estimate.
- What happens when a site changes its layout?
- The markup stops matching and the extension says so on the first page rather than filling empty cells. Usually you re-mark it by the same selecting, and a repair is inside support.