October 8, 2026
Reddit Scraper gains Sort Period; Facebook Marketplace Scraper gains Newer than, an option to exclude results outside your search, and a new is_outside_search field; a blocked or down website now pauses the run instead of ending it in error; Pinterest pins with an encoded id no longer stop the run
Reddit Scraper gains Sort Period; Facebook Marketplace Scraper gains Newer than, an option to exclude results outside your search, and a new is_outside_search field; a blocked or down website now pauses the run instead of ending it in error; Pinterest pins with an encoded id no longer stop the run
New
Reddit Scraper — Sort Period
sort_periodsets the time window of a ranking:all(default),year,month,week,dayorhour. It applies to subreddit feeds sorted bytoporcontroversial, and to searches sorted byrelevance,toporcomments. Atalready in the task URL (e.g.https://www.reddit.com/r/Bitcoin/top/?t=year) wins.- Squids already set to
toporcontroversialnow rank over all time. Reddit's own default for these feeds is the last 24 hours, which is what they returned until now: setsort_periodtodayto keep that.
See update-settings.
Facebook Marketplace Scraper — Newer than and Exclude Results Outside Your Search
fetch_since(Newer than) keeps only listings published after a point in time: a relative duration ("24h","7d","2w") or a date (YYYY-MM-DD, read infetch_since_timezonewhen set). On search URLs, Facebook's matching Date listed filter (last 24 hours, 7 days or 30 days) is added for you, and the older listings Facebook still returns are skipped one by one and not billed. Set it to your schedule interval to collect, and pay for, only the new listings of each run.exclude_outside_search(off by default) skips the listings Facebook shows under "Results from outside your search" (outside your search area, or only loosely related to your query) and stops paging once that section starts. Skipped listings are not billed.- New result field
is_outside_search:truefor the listings under that section,falsebefore it, empty on browse and category URLs.
See update-settings and get-results.
Improved
Reddit Scraper — feed limits spelled out
Reddit serves about 1,000 posts per feed at most (new, top, controversial, search), so on a busy subreddit new only reaches back a few weeks. The update-settings page now explains how to go further back (several sorts and periods, plus keyword searches scoped to the subreddit), and that fetch_since ends the task at the first older post, so it belongs with sort_subreddit set to new.
Facebook Marketplace Scraper — Get Listing Details cost spelled out
Get Listing Details is billed once per squid: on later runs, the listings this squid already enriched are skipped and not billed again. This was already the behaviour; the credit cost page now says so.
All scrapers — a blocked or down website pauses the run instead of ending it in error
When the website being scraped refused our requests (HTTP 403 or 429) or did not answer properly (HTTP 5xx) even after retries, the run used to end as ERROR. It now pauses for 10 minutes with done_reason website_inaccessible and resumes on its own, picking up only the unfinished tasks, up to 3 times per run; a fourth time it ends as ERROR as before. Runs you poll or receive by webhook will show PAUSED then RUNNING again instead of ERROR. See errors.
Fixed
Pinterest Scraper — pins with an encoded id
Pinterest now gives some pins an encoded id such as e1_AfqPShzqO1C7_LEkrAbtoIWUe2MoGdDTj5vU32Lr9gXn25qs1FpIvg instead of a number. These pins made the run stop with an error; they are now collected like any other pin. pin_id stays a string: store it as text, not as a number. See get-results.