Scrape X (Twitter) replies, comments & whole conversations from $0.00015 per delivered reply. Read Twitter replies without an account. Paste post URLs in bulk, use 25+ filters & export flat rows. Built by Xquik. World's fastest & cheapest X (Twitter) scraper service. Not affiliated with X Corp.
2026-10-24. Each photo in media carries editedWithNativeEditor: true when
its author edited it in X's own editor.
2026-10-21. Large runs & runs with a problem save run-report again. It
counts failed targets by cause in targetFailureKinds.
2026-10-21. softIntervention holds the label X shows on a tweet it limits.
2026-10-17. A run now ends as soon as its work is done, even while Xquik is
down.
2026-10-18. A run for all replies of a post with more than 300 now saves every
nested reply too, up to maxItems. Before, it saved nested replies only
within its first 300 rows.
2026-10-18. A run now says X was unavailable when X's outage stops a target.
2026-10-17. Each post lists the counts X did not send in unknownCounts.
Those counts read 0, so treat them as unknown. Rows in snake_case name them
like their fields, such as bookmark_count.
2026-10-14. A profile link by user ID, like x.com/i/user/11348282, now reads
that account.
2026-10-14. A request that gets no answer is sent again beside it, so its
replies arrive sooner.
2026-10-14. A run on a post older than 7 days says X shows only part of its
older replies, & no longer that X hides them.
2026-10-14. The README says the prefilled form holds a conversation we tested.
2026-10-14. The Store text & the top of the README use plainer sentences.
2026-10-13. Each post's author now holds profilePictureOriginal, the profile
picture at full size, to download. profilePicture stays the 48 px image.
2026-10-13. An account can be named by its user ID of 16 to 19 digits, like a
handle. This works in usernames & profile links.
2026-10-13. A run with scopedirect or collectionStrategyreplies says
how many replies X hides. Before, an empty one said no replies matched.
2026-10-13. New example tasks collect giveaway replies & read Twitter comments
without an account.
2026-10-13. The Store title reads X (Twitter) Replies Scraper, & the README
opens with Twitter replies & comments.
2026-10-13. Field help & the README drop colons & semicolons.
2026-10-13. The run status drops its colons. An aborted run starts Aborted.,
& the batching tip has no label.
2026-10-13. An empty run of a post whose replies X hides now says so in its
status. Before, it said no replies matched.
2026-10-12. A list that keeps starting over from its first page ends after 2
restarts in a row, where it read that page again until the run's time ran out.
No restart charges a row twice.
2026-10-12. sortBy, queryType, includeNestedReplies,
includeRepliesOfReplies & outputVariant take effect when sort, scope &
outputMode stay at their defaults. Before, the default won.
2026-10-12. The Actor reads more input names of other reply Actors, which are
resultsPerCategory, includeSourceTweet, excludeOpSelfReplies, depth,
onlyDirectReplies, twitterUrl & the sort words top, newest &
mostLiked. A category that is another Actor's job stops the run before any
read or charge.
2026-10-12. Shape your rows with fields, omitFields, renameFields,
flatten, joinFields & constantFields. Name totalEngagement,
engagementRate or engagementPerView in fields to add it.
2026-10-12. Authors carry email & contactInfo, the emails, phones & links
their profile publishes, when X sends the bio.
2026-10-12. subscriberBadge names the creator whose Subscriber badge a reply
shows.
2026-10-11. Limited likes runs of direct replies no longer miss top replies
on large threads.
2026-10-11. The schema says X sends source, the posting app, for some posts
only.
2026-10-11. A run whose posts all ended no longer writes an empty
next-cursors, so runs cost less.
2026-10-10. Replies carry cashtagAttachments, the price charts their authors
attached to cashtags. Tags in entities, hashtags included, add their
SmartTag id & active when X sends them.
2026-10-09. Each Grok image in media carries grokImageAnnotation, the
prompt it came from, when X sends one.
2026-10-09. Runs of all or nested replies by relevance or latest also save
each page as it arrives, with each reply's depth.
2026-10-09. Runs of direct replies by relevance or latest save each page as it
arrives. A stopped run keeps what it read.
2026-10-09. Unfiltered runs of direct replies sorted by likes finish sooner
with a limit.
2026-10-09. Filtered runs of direct replies finish sooner.
2026-10-09. A filtered run's status counts the replies it read, not only the
ones that matched your filters.
2026-10-09. Set mediaType to none to keep only text replies, without media
or links.
2026-10-09. Successful runs with rows end their status with a short review
request.
2026-10-09. The default auto strategy adds the replies the post's author
hid, as hiddenReply rows. Set includeHiddenReplies: false to skip them.
2026-10-08. Long threads return more replies. When the first reads cover part
of a thread, the run reads the rest of its replies too.
2026-10-08. The Actor skips post IDs X cannot have, such as 20-digit ones, &
names them.
2026-10-08. Replies sorted by likes or oldest first now read in larger pages,
so long threads finish faster. You get the same replies in the same order.
2026-10-07. A read that gets no answer retries at once, & a stalled lookup
gives up after a few seconds, so runs lose less time to slow requests.
2026-10-07. Community tweets can carry communityName beside communityId
when X sends the community's name.
2026-10-07. A run that reads every target & finds no reply writes no
diagnostics record. Its status says why, & the run costs less.
2026-10-05. Runs with many targets save progress less often, finish sooner &
cost less.
2026-10-05. Large runs upload their rows faster.
2026-10-05. Runs sorted by most liked now find liked replies that X's ranking
skips.
2026-10-05. Runs of under 5 IDs or handles end with a tip to batch them.
2026-10-04. A run on a renamed handle now saves the replies X shows for it.
2026-10-04. A run whose posts all ended writes next-cursors only when it
keeps its report, so small runs cost less.
2026-10-04. The Actor version in diagnostics & run reports is now the Apify
build number.
2026-10-04. A run no longer stops early when an upload's answer is lost. Its
rows are charged once.
1.2.125, 2026-10-04
collectionStrategy: "hiddenReplies" returns the replies a post's author hid.
Each one is a hiddenReply row at the usual row price.
1.2.124, 2026-10-04
Every run uses 256 MB again, which is faster & more reliable for small runs.
1.2.123, 2026-10-01
When Xquik stays busy, each retry waits longer than the last.
If Xquik stays busy too long, that target fails & the run says why.
Busy answers are never charged.
1.2.122, 2026-09-30
A run with rich output keeps its run-report.
1.2.121, 2026-09-30
A run near its time limit stops before a read it can't finish.
1.2.120, 2026-09-29
A run that started late says how long it waited, out of its time limit.
1.2.119, 2026-09-29
Requests to Xquik never follow a redirect, so the key stays with Xquik.
1.2.118, 2026-09-29
A run waits out a brief Xquik outage instead of failing.
Rows saved before an outage stay saved & are charged once.
1.2.117, 2026-09-29
Runs sorted by likes or oldest, or with nested scope, keep 256 MB.
Runs with more than 5 targets keep 256 MB, whatever field names them.
1.2.116, 2026-09-29
Runs up to maxItems 50 with up to 5 targets use 128 MB & cost less.
1.2.115, 2026-09-29
New outputMode: "rich" keeps every full column & adds hasCommunityNotes,
quotedTweetPermalink, unavailableAuthorId & media sourceUser.
Full, the default, & every other output keep their rows unchanged.
1.2.114, 2026-09-29
The README names the Actor as Xquik's X Reply Scraper & adds a how-to & FAQ.
The README says the prefilled form returns up to 1,000 rows, as the form sets.
1.2.113, 2026-09-29
Runs sorted by most liked stop once X's most liked replies fill the limit.
Runs with a reply author skip a second read after the first one finished.
1.2.112, 2026-09-28
A run that finds nothing skips run-report, so it costs less in Apify usage.
1.2.111, 2026-09-28
Runs under 34 results cost less in Apify usage, with no progress saves.
1.2.110, 2026-09-28
An aborted run's status counts its saved results, even during a slow upload.
1.2.109, 2026-09-28
A missing post or account no longer counts as a failure. The status names it &
says to check the handle, ID or URL.
1.2.108, 2026-09-28
Posts from 2006 to 2010 now get their replies, from a URL or an ID.
1.2.107, 2026-09-27
During a run, the status shows how many results it has collected so far.
The status says when the run waits to retry.
1.2.106, 2026-09-27
The new alwaysSaveRunRecords input keeps run-report on every run.
1.2.103, 2026-09-27
URLs without https://, such as x.com/nasa, now work.
Post URLs in capitals & Article URLs now read that post's replies.
A post URL with a broken ID no longer reads its author's replies.
Profile URLs whose handle is all digits, such as x.com/1234567, now work.
The run lists each input it cannot read & says how to fix it.
With nothing to read, the run writes a free diagnostic with the fix.
1.2.99, 2026-09-27
Runs open their dataset while they read X, so they finish sooner.
1.2.96, 2026-09-27
Small runs that go well skip run-report, so they cost less in Apify usage.
The run status now says why the run stopped, with its charged results &
targets.
Runs that stop on their input now name the diagnostics output in their status.
1.2.95, 2026-09-27
Runs that stop for missing or unusable input cost less in Apify usage.
1.2.91, 2026-09-27
Support now sees a finished run even while the service is busy.
Runs with a time limit keep a little more time to finish.
1.2.90, 2026-09-27
Runs with a time limit now stop in time & keep their results.
Large runs save their progress faster.
1.2.89, 2026-09-27
The run status now says where to find its diagnostics.
1.2.87, 2026-09-27
If the Actor cannot start, the run now ends with 1 free diagnostic.
Runs with hundreds of targets use less CPU.
1.2.86, 2026-09-26
Runs with a date range skip recovery reads that found nothing.
A date range with no replies now ends empty, not as a partial run.
1.2.85, 2026-09-26
Runs now start collecting results sooner, so small runs finish faster.
1.2.84, 2026-09-26
Filling maxItems no longer logs errors for targets the run did not need.
A run that fills maxItems no longer reports a time limit.
1.2.83, 2026-09-26
Large runs use less memory.
1.2.82, 2026-09-26
Targets no longer fail when Xquik is busy. The Actor waits & tries again until
its time limit.
A page the service answers slowly in busy hours now arrives instead of timing
out.
1.2.81, 2026-09-26
Support can now see where each run spends its time.
1.2.80, 2026-09-26
Targets no longer fail when the service is busy. The Actor waits a few seconds
& tries again until its time limit.
1.2.77, 2026-09-25
Input & Output tab descriptions now say what each field does.
The README now focuses on what you get & how to run it.
Runs finish sooner when every reply to a post is already collected.
1.2.76, 2026-09-25
Support can now look up every finished run.
1.2.75, 2026-09-25
Runs with many posts or profiles finish faster.
1.2.74, 2026-09-25 update
Small reply threads finish much faster.
The status says how many replies X hides. It no longer asks for a rerun.
A run with a time window ends as complete once it reads that window.
1.2.74, 2026-09-25
Runs recover faster from temporary service delays.
A slow page now ends in time & keeps its rows.
1.2.73, 2026-09-25 update
A post that cannot be fully read keeps its rows & ends as partial.
Runs wait the right time before retrying an empty answer.
Partial posts no longer count as failed targets or log errors.
Languages id, he, fil & nb now keep their replies.
1.2.73, 2026-09-23 update
The status now names every reason a run stopped early.
The diagnostic lists each cause in stopCauses, with advice.
1.2.73, 2026-09-20 update
Brief stalls on the first page no longer end a run as incomplete.
Large & old threads read much faster.
1.2.73, 2026-09-17
The status shows how much of a large thread X made available.
Runs end as complete once every reply in range is collected.
1.2.72, 2026-09-17
Runs report completion more accurately.
1.2.71, 2026-09-17
Runs finish sooner once every reply is collected.
Runs recognize posts that have no replies.
1.2.70, 2026-09-16
Find replies to posts inside larger conversations.
Unrelated conversation branches stay out of the results.
