What Each X Article Export Format Actually Keeps (and What It Drops)

Long-form posts on X are the format people lose first. The page is built to be scrolled, not archived, and a screenshot cuts a 3,000-word piece into unreadable slices. So the useful question is not "which saver is best" but "what does each export format actually keep".

That question has a measurable answer. Below are the numbers from one two-post thread exported through TweetArticleSave in four formats on the same day, plus the two input mistakes that break the export before format even matters.

The input mistake that comes first

A thread is not one post. If you paste the link to the first post, you get the first post: 891 characters in one measured case, 381 in another. The rest of the thread never arrives, because the first post does not contain it.

Paste the link to the last post instead. The same two threads then returned 3,672 and 848 characters with the full chain intact. Nothing in the interface tells you this; it is simply how the page is structured, and it is the single most common reason an export looks short.

Four formats, one source document

Source: a two-post thread, 3,672 characters, one embedded image.

FormatFile sizeWhat you get
PDF548,977 B3 pages, text and image laid out
DOCX443,250 BEditable text; the image is embedded inside the file (418,440 B of the total sits in word/media)
Markdown (ZIP)405,019 BA ZIP containing a 4,334-byte .md file, a SOURCE.txt, and an images/ folder
TXT4,149 B3,456 characters — every text block, images as filename placeholders

Read the last two rows side by side. Until a code change on 23 October 2026 the TXT path wrote 1,212 characters of this 3,672-character source and silently skipped both quoted-post blocks (286 and 2,624 characters). Re-measured on 27 October 2026 against the same thread, TXT carries all 12 text blocks and 3,456 characters, quoted blocks included — but each image is replaced by a filename placeholder such as [image: 01.jpg]. TXT is usable for text and useless for pictures. If you need the image, take the DOCX or the Markdown ZIP.

So which one?

  • Reading or printing: PDF. Three pages, no reflow surprises, image in place.
  • Editing or quoting later: DOCX. It is the only format where the image is embedded as a real object you can move.
  • Version control or static site: Markdown ZIP. The .md is small (4,334 bytes) because it references images by path rather than inlining them — which is what you want for a repo, and a broken export if you only copy the .md and forget the folder.
  • Plain text: TXT is fine for words and drops images. If you need both, export DOCX and strip formatting locally.

What is not on the list

There is no image export. The formats are DOCX, PDF, Markdown, TXT and a ZIP bundle of the Markdown set — no option that renders the article as a single picture. If a guide tells you to "export as image", it is describing a screenshot with extra steps.

Two questions

Does an account-free export still get the whole article? Yes — the account is not what limits the body, the link you paste is.

Does this work on a single long post rather than a thread? A single post is the simpler case: paste it, and the chain logic does not apply.

Is length ever truncated? The measured sample was 3,672 characters and came through whole; that is one sample, not a guarantee for arbitrary lengths. What TXT drops is not tied to length: since 23 October 2026 it keeps every text block and drops the images.

The whole job is choosing a format that keeps what you need and knowing which one quietly does not. Four formats, four different answers — and the TXT one still leaves the pictures behind.

Disclosure: I work on TweetArticleSave (tweetarticlesave.com), the tool used for these exports, so treat the descriptions here as a vendor statement rather than an independent review. All figures come from the same two-post thread exported in four formats on 21 October 2026. This post replaces an earlier version that promised embedded media and no length limit without distinguishing formats; the TXT limitation above was documented the same day; the quoted-block gap it described was closed in the code on 23 October 2026 and the figures above were re-measured on 27 October 2026.

评论

此博客中的热门博文

Dailymotion Videos That Will Not Save in 2026: The Three Reasons

The AI Human Pyramid Video on Facebook: Why Physics Says It Never Happened

Why Some Instagram Reels Save and Others Do Not (2026)