Scrape fiverr.com gigs at $0.75 per 1,000 — starting price, every package tier with delivery time and revisions, rating, review count, category, seller level, and whether a listing is promoted or organic. Search by keyword or paste a Fiverr search URL.
Incremental runs on the same state key no longer have to be 30 minutes
apart. Before, a run that started soon after the previous one finished
exited with 0 results and the message "Another run is already updating
this state".
A run that the platform restarts midway now continues where it stopped:
gigs already saved are not fetched, stored or charged again, and search
pages already read are skipped. Before, the restarted run started over
from the first page and could stall for minutes.
When only some gigs get full details, the run log now says how many
("Full details loaded for 34 of 40 gigs") instead of showing an error.
An error is still shown when no gig on a page gets details.
More gigs now come back with full details (description, packages, gallery).
Fiverr blocks a larger share of gig page requests than before, so each gig
page is now tried up to 16 times instead of 8. Detail fields remain
best-effort: a gig whose page stays blocked keeps detailFetched: false.
Gig URLs that Fiverr briefly refuses are tried again after a short pause
instead of being given up, so a run with only gig URLs no longer fails with
0 results when Fiverr is busy. If a gig still cannot be loaded, the run now
says the gig page was refused rather than "search page 1 unreachable".
Several searches in one run: the new Search keywords (several) list takes one
keyword per line; Max Results is the total across all searches.
Single gig URLs (https://www.fiverr.com/ /) are now accepted
in Start URLs and returned as full rows, and can be combined with the keyword
list in one run. Seller and category pages are still
skipped with a warning.
With gig details on, every row now also carries the seller's country,
languages, member-since date and seller-wide rating and rating count.
searchQuery now shows the keyword of the search that found each gig; in runs
with several searches every row used to carry the first keyword.
Gig details restored: description, packages, gallery, rating and orders in
queue are loaded from the public gig page again (the app detail route started
requiring device attestation). publishedAt is no longer available from Fiverr
and stays empty.
When full gig details cannot be loaded, the run log now says so instead of
finishing silently. Affected rows keep every field; the detail fields
(description, packages, gallery) are empty for those rows.
Fixed searches that failed with "search page 1 unreachable" after about 100
seconds when Fiverr was briefly slow to answer: retries and fresh connections
now actually reach Fiverr again instead of giving up immediately.
Fewer gigs without full details when Fiverr is busy: a gig page that is
briefly refused is tried again a few more times, with longer pauses, before
its detail fields are left empty.
0.1.0 — unreleased
First release.
Extracts Fiverr gigs by search keyword or Fiverr search URL: title, starting
price, rating, review count, category, seller (username, display name, level,
online status) and placement (promoted vs organic, rank within the result set).
Optional detail enrichment adds the full description, every package tier with
price, delivery time, revisions and features, the media gallery, publish date
and current orders in queue.
Incremental mode reports NEW / UPDATED / UNCHANGED between runs. Ranking,
rating and review-count churn deliberately do not count as changes — they move
on almost every gig between runs, and treating them as content changes would
re-report the whole result set every time.
EXPIRED is reported only when a search's full result set fits inside the
platform's own result window (~1,000 results per query). Broader searches
cannot observe the whole set, so nothing is reported as expired rather than
guessing that unseen gigs are gone.