How to use Cursor to extract data from HTML
A developer used a worksheet and case-specific notes to guide Cursor through repeated HTML extraction tasks, reporting substantially less effort while retaining manual cleanup.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Use [scraping worksheet] and [case-specific notes] to extract [required fields] from [HTML input]. Filter to the relevant content, implement the extraction, and write tests. Identify anything that needs my cleanup or review.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- iambateman
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reported fairly successful extraction, good tests, and a 75% reduction in the energy required for a tedious task, with manual cleanup still necessary.
- Human involvement
- The author supplied instructions for each case and explicitly reported cleaning up after Cursor finished.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Write a worksheet describing the recurring scraping task.
- Give Cursor specific notes for the current extraction case.
- Let Cursor filter the HTML and parse the required data.
The author described a repetitive development task involving scraping a site for similar pieces of data roughly 200 times. They prepared a worksheet for Cursor, then supplied specific notes explaining how to handle each particular case. That combination gave the agent both a recurring task description and instructions tailored to the current extraction.
After receiving those materials, Cursor ran the task. The author found it fairly successful at filtering HTML down to the relevant content and parsing out the data they needed. They also reported that it wrote good tests. The account does not include the worksheet, an example extraction, or the contents of those tests.
Human cleanup remained part of the process: the author explicitly said this was never truly hands off. Their reported benefit was that Cursor removed 75% of the energy from work they found particularly tedious, rather than eliminating their involvement. Who it's for: developers handling repeated HTML extraction tasks who can provide case-specific instructions and inspect the resulting work.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · iambateman
Frequently asked questions
What instructions did the developer give Cursor?
They wrote a worksheet and supplied specific notes explaining how to accomplish the scraping task in each particular case. The actual worksheet text was not included.
Did Cursor complete the scraping without human involvement?
No. The author said it was never truly hands off and that they still needed to clean things up after Cursor finished.
More details
Remaining steps
- Clean up the work after Cursor finishes the task.
About this account
Cursor ran individual extraction tasks from a worksheet and case-specific notes, but the author still performed cleanup.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=43309999. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: iambateman on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Cursor
Web scraping · HTML · Data extraction
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.