For replies, look for a reader that opens the surrounding conversation. For a thread, check that every required post is present. For an image or a raw timestamp, a smaller media or data view may finish the task faster.
The ten illustrated entries record a logged-out desktop pass on July 30–31, 2026, with several flows repeated on August 1. An October 15 check below follows the same thread and photo inputs through actual returned content and Markdown. The older screenshots keep their original context.
CarryFeed is our product. Its entry covers the public-post-to-AI handoff observed in the same pass.
Contents
Choose by the result you need
- One post for an AI question
- CarryFeed observationThe result kept text, photos, author, date and the original link together, then placed its result URL in an editable ChatGPT composer.
- Replies around a post
- TwitterWebViewer observationThe status URL returned the post and a reply list.
- One photo from a set
- Xhancer observationBoth photos had separate controls; compare the saved file with the media row.
- Raw fields from a status ID
- xAPIs.dev observationThe numeric status ID returned a readable card with raw timestamp and URL fields.
- Profile browsing
- TweetViewer.comTwitterViewer.netXTwitterViewerHandles opened profile views in the initial pass. TwitterViewer.net later stopped at human verification.
- Quoted context or a full thread
- viewXXunrollThe historical results differed on quote context. The October comparison below also separates a 19-entry sequence from three additional replies; a larger count alone is not a better result.
- An editable document
- Xtracticle observationXtracticle’s October opener input produced a 19-section Markdown document with quote context. Check source links and media in the actual file before using it.
A thread count can hide a different selection of posts
On October 15, the same Julia Bergeron opener produced 22 body entries in Xunroll and 19 numbered sections in Xtracticle’s Markdown. Comparing the content explained more than comparing the totals.
The first 19 Xunroll entries followed the launch-viewing sequence through the refuge-road closure update. Its last three entries were additional replies addressed to @Eric_Larkin, @T_De_La_Rosa and @g8trdavid. Xtracticle ended at that closure update. Its document also expanded quoted posts that Xunroll’s body omitted.
- Xunroll: 22 body entries, broader reply selection
- Open the observed thread pageThe main sequence and three additional addressed replies appeared together. Twenty linked images appeared in that body. The Brevard EOC quote text was absent, and both View on X links still used the misspelled handle ulia_bergeron. Keep the source URL you supplied.
- Xtracticle: 19 numbered entries, expanded quote context
- Open the same input in XtracticleThe opener now produced a multi-post document without asking for the final URL. The actual Markdown contained the opener, the closure update at 19/19, 20 main-post image links and four quoted-image links. The Brevard EOC author, text and original status URL remained separate from Julia’s comment.
- Two-photo post: verify the files listed in the export
- Original Mbappé sourceThe Xtracticle Markdown kept @KMbappe, July 5, 2026, the exact source URL and both photo URLs. The two media IDs matched FxTwitter’s public post response in the same order. These were image links in Markdown, not an offline image bundle.
For this launch-viewing sequence, Xtracticle supplied a more useful working document when the quoted notices mattered. Xunroll exposed additional conversation replies. Choose that scope deliberately; 22 versus 19 does not by itself establish which export is complete.
The thread Markdown identified the opening source and quoted sources but did not attach a permalink to each main-sequence entry. For a citation to one later post, obtain that post’s individual URL. This check matched the returned sequence’s first and last text; it did not establish exhaustive coverage of every current X reply.
The content check used Xtracticle’s publicly documented Markdown endpoint for the opener and the two-photo post, alongside the visible browser results. The file’s timezone-free date display is not a substitute for an exact timestamp when chronology matters.
If you came here looking for a Sotwe alternative
Choose the task you want to continue: opening a known profile or post, finding an unknown source, following trends, or saving media. Each needs a different starting point.
For a saved thread, start with the completed content comparison above. If you need quoted notices, compare the quote text and source URL in the document. If you want surrounding replies, inspect which replies the reader included; an extra reply can inflate the count without extending the main sequence.
- You have a public status or profile URL
- CarryFeed Twitter ViewerPaste the exact URL. Keep the original X link with the returned text and source details. For a thread document, the October Xtracticle result above provides a concrete alternative; its 19 sections and quoted notices were inspected, with no claim to every reply.
- You remember a name, phrase or date
- Account-and-clue search guideAdvanced SearchFind the status URL first. A profile reader cannot identify a post from a vague memory.
- You want the attached file
- ImagesVideoOpen the appropriate media tool with the original post URL. Match each returned file to the post and check its dimensions.
- You were browsing topics
- Twitter TrendsStart with a location’s topic list, then open the associated search before choosing an individual source to read.
- You want RSS or a Nitter-style front end
- Nitter migration guideCompare feed subscriptions, self-hosting and mirror access as separate needs.
For public reading, use a service that does not ask for your X password. Protected posts require access through X; switching reader domains does not change their audience.
The same sources made the differences visible
The initial pass used Kylian Mbappé’s profile and a two-photo status wherever the tool accepted them. Julia Bergeron’s Artemis II thread supplied a second input for quote context, thread reading and exports. The August 1 repeat exposed several changed or partial results recorded below.
These checks captured results and friction. They did not measure uptime, speed, network privacy or long-term free allowances.
- Profile
- @KMbappeAn exact handle for profile readers.
- Two-photo post
- Status 2073559834096492936A shared status URL for post, image and numeric-ID inputs.
- Thread opener
- Julia Bergeron’s Artemis II threadThe shared opening URL used for thread and export checks.
1. CarryFeed
Observed workflow: public post to an editable AI question, with the source retained.
The Mbappé result kept the caption, both match photos, author, date and original X link together. Its result URL then moved into an editable AI prompt.
The next step begins with an editable question
Clicking ChatGPT opened a new conversation with the CarryFeed result URL in the composer. I added a question about how 'au boulot' sounded beside the match photos, then checked the draft before sending it.
Open a public post in CarryFeed
2. TwitterWebViewer
Observed workflow: post and replies, then the author’s profile.
The Mbappé status URL opened as a complete post, with his caption and both match photos above the engagement figures. Copy link and Bookmark sat above the reply list.
I read the replies to see how people reacted to the match photos, then used the handle to open the account timeline.
Choose it when the conversation matters
Searching @KMbappe opened the profile view, where the banner and bio led into the public timeline. Media stayed available from the same header.
3. Xhancer
Observed workflow: separate file controls for each photograph.
Xhancer opened the status as a single post card. Mbappé's caption stayed above both photos, with the original link nearby.
Both frames stayed visible, and the second photo had its own Download and Copy URL controls.
Each image gets a separate decision
The page stayed focused on the media, so I could choose a single file without opening a profile feed.
4. xAPIs.dev
Observed workflow: numeric status ID to a readable card and raw fields.
xAPIs.dev asks for the numeric status ID, so I copied 2073559834096492936 from the Mbappé URL into the field. Comma-separated IDs can share the same request.
A readable card appeared above the raw response, showing the author, caption, engagement figures, and a View on X link.
Use the card first, then inspect the fields
The JSON included the exact timestamp and x.com and twitter.com versions of the post URL.
5. TweetViewer.com
Observed workflow: handle to profile; the direct status downloader returned a partial photo set.
Searching @KMbappe opened the profile with recent posts under the bio. Export stayed beside the feed tabs, and the visible image had its own download control.
Profile browsing leads into a separate downloader
I started with @KMbappe and found the image post in the profile. Its status URL opened the media downloader, where one photo filled the result and @KMbappe appeared below it.
6. TwitterViewer.net
A tabbed profile in July; the August 1 repeat stopped at human verification.
On July 30, the Profile Viewer opened Mbappé's banner and bio above the recent feed. I used Media for photos, then opened Replies from the same account page.
The account tabs stay within reach
After pasting the handle, I moved from Media into Replies from the same header. Both tabs stayed in view as I browsed.
7. XTwitterViewer
Observed workflow: profile or post in a compact reading column.
Tweet mode opened Mbappé's status as one card, with his caption above both photographs and the engagement figures nearby. On desktop, the centered column felt close to reading a post on a phone.
Profile mode uses the same narrow reader
I selected Profile on the home screen and entered @KMbappe. The account timeline opened with his Instagram link and joined date under the header.
8. viewX
Observed workflow: one post with its quoted notice visible.
Julia Bergeron's Artemis II update kept the quoted Brevard EOC notice inside the reading view. Its April 1 launch window sat beside Julia's traffic advice.
The quoted notice stays in context
Image and PDF controls appeared above the reading view. The 22-post count appeared after I reused the same URL in Xunroll.
9. Xunroll
Recorded workflow: same-author sequence and conversation replies on one page.
On August 1, Julia’s opening URL returned a page headed 1/22. Scrolling revealed 22 same-author entries on that page; the quoted Brevard EOC post visible in viewX was absent from its body. The October content check above distinguishes the main sequence from three addressed replies.
The body matters more than the 1/22 header
Save as PDF, Copy thread content, Reading mode, and Hide media appeared with the thread. Previous and next controls remained in the header, but they were not required to read the full sequence on the page.
10. Xtracticle
July/August: an opener-only export; October: a 19-entry document with quoted context.
I pasted Julia's opening status URL into Xtracticle and chose Markdown from the result.
The earlier page returned Julia’s opening post and said a complete thread needed the final post URL. That July/August observation is preserved here; the October content comparison above records the changed result.
Markdown becomes the working copy
Markdown, text, and PDF appeared as separate downloads above the extracted post. Copy and Share stayed beside them.
Check the piece you need before copying it
- Replies
- Open one reply by its direct link and find it in the returned conversation. Record a missing reply as a gap in this result.
- Thread
- Compare the first and final posts and inspect the intervening sequence. A 1/22 header alone does not establish that all 22 posts were returned.
- Quote
- Check whether the quoted post’s author, text and original URL remain separate from the quoting author’s comment.
- Media
- Check that every required image has its own file or source row. A successful lookup can still return only part of a photo set.
- Export
- Open the downloaded document and compare it with the source. Keep the original URL alongside the copy.
Add the question before opening an AI conversation
In the recorded CarryFeed flow, ChatGPT opened with a result URL in the composer. Add the wording, image or source detail you want to examine before sending.
Explain how “Une journée au boulot” changes tone beside the match photos. Use the post text and photos supplied here, and retain the author and original X URL.
If the result page is unreadable in your AI client, use the source-copy workflow to provide the text and media directly.
Try the reader with your own URL
Choose the entry whose recorded result matches your task, then test the exact post or profile you need. Check missing fields before moving into replies, media or an export.
- Open a known public source
- Open CarryFeed’s Twitter ViewerUse the full status URL or exact handle. For a thread, verify that the returned sequence contains the posts you plan to use.