Kickstarter Scraper

Crowdfunding projects from Kickstarter as a clean table. Search by term, filter by state and category, and get one row per project with the funding goal, the amount pledged, how far past or short of target it is, the backer count, the deadline and who is running it.

Leave the search term empty to browse rather than search, and use Sort by to get the view you actually want: most funded, most backed, ending soonest, or newest.

Money is reported twice, on purpose

Kickstarter runs in many currencies, and this is the detail that quietly ruins an analysis. goal and pledged are in the project's own currency, which is why currency sits next to them. pledged_converted is the same figure converted, with converted_currency naming what it was converted to.

Sort a mixed set on pledged and you are comparing euros with yen without being told. Sort on pledged_converted and you are not. Both are here so the choice is yours and visible.

percent_funded is a percentage, not a fraction: a project at 930% raised nine times its goal. Kickstarter's own payload sends 9.3 for that, which is easy to read as nine per cent.

pledged_per_backer is worked out for you, because average pledge is the first thing anyone computes from this data and it is null rather than a crash when a project has no backers yet.

What you get per project

• name, blurb, url, state
• goal, pledged, currency, pledged_converted, converted_currency
• percent_funded, backers_count, pledged_per_backer
• category, parent_category, category_id
• creator_name, creator_id
• country, country_name, location
• created_at, launched_at, deadline
• is_staff_pick, is_spotlight, photo_url
• search, search_page, position, total_hits

category_id is worth keeping. Run once without a category, find the id of the one you care about on any row, then use Category id to narrow every run after that.

Run sizes

Projects arrive twelve to a page and the endpoint says when it has run out, so a run stops on its own rather than walking to a page limit. Maximum projects caps the total and Maximum pages per term is the rail on a broad search.

Errors

Code	Meaning
no_results	The search ran and matched nothing
fetch_failed	The endpoint could not be read after several attempts

no_results and fetch_failed are deliberately different. A term nobody has launched a project for and a request that was refused look identical if you only count rows, and only one of them is worth rerunning.

A note on reliability

Route access here is narrow and can change without warning. If a run suddenly returns fetch_failed across the board where it worked last week, that is what has happened; it is not a bad search term. Re-run once, and if it persists, say so rather than working around it.

Related actors

For launches on the other big platform, use Product Hunt Launches. For the companies behind a project rather than the project itself, look at the LinkedIn company actors.