Open TweetArticleSave

Why an Exported X Thread Comes Out Incomplete

Updated 2026-10-27 · two public test threads, exported and opened file by file on this date

An export that stops early is rarely a thread that was too long. Here are the five reasons we could reproduce, what each one looks like in the file, and the fix that actually works.

An export that looks cut short is usually not a thread that was too long. On 2026-10-27 we ran the same two public threads through this site's own parser and downloaded all four formats, then opened each file and counted what survived. A 2-post thread of 3,672 characters came out complete — all 12 content blocks, the image, and both quoted posts — in Word, PDF, Markdown and plain text.

When something really is missing, it is one of five causes. Each has a different fix, so it is worth identifying yours before exporting again.

  1. You pasted the first post of the thread. The chain is rebuilt upward from the link you give it, so the first post carries only itself.
  2. You chose plain text and expected pictures. TXT keeps the words and replaces each image with a placeholder.
  3. The tab was open for more than three hours. Parsed documents are dropped, and the download refuses with a message about the source.
  4. The post cannot be read while logged out. Protected, deleted and paywalled posts return an error instead of a partial file.
  5. You expected a quoted post to arrive as a copy. It arrives as one bracketed note, not as the quoted thread's other posts.

1. The first post carries only the first post

This is the most common reason a thread looks truncated, and the most misleading, because nothing fails: you get a file, it is just short. Paste the permalink of the last reply and the tool walks backwards through the reply chain. Paste the first post and it has nothing above it to walk to.

Same author, same two threads, four links, read on 2026-10-27 through this site's parser. Character counts are what the parser reported.
Link you pasteResultPostsCharacters
Thread A — first post
…/status/2109990238533677520
single post1891
Thread A — last post
…/status/2111971887811342582
thread23,672
Thread B — first post
…/status/2111914776217747673
single post1381
Thread B — last post
…/status/2111938337607172591
thread2848

The four numbers match a second run of the same links on 2026-10-20 exactly, so this is behaviour, not a one-off. If you already exported the first post and the file is on your disk, re-read the last reply and export again — there is no way to extend an existing file.

2. Each format drops something different, and TXT drops the pictures

Below is the same thread (A, 3,672 characters, 12 content blocks, one image, two quoted posts) in all four downloadable formats. "Blocks kept" counts the paragraphs, list items, post headers and quoted-post notes that were present in the parsed document and found again in the exported file.

Thread A, exported and inspected 2026-10-27. Sizes are the files the site returned.
FormatWhat you getSizeBlocks keptImages
Word .docxone editable document, header and footer included443,251 B12/121
PDF .pdfa 3-page document with a real text layer (3,517 characters extractable)548,977 B12/121
Markdown .zipa ZIP holding the .md, a SOURCE.txt and an images/ folder405,019 B12/121
Plain text .txtcharacters only, no styling, no pictures4,149 B12/120

Two things follow from that table. First, TXT is not a broken export: it kept all 12 blocks, including both quoted posts as [note] [quoted @handle] lines. It simply has no images — each one becomes a [image: 01.jpg] placeholder, which is exactly what makes a plain-text file look thin next to a 3,672-character thread with pictures.

Second, a PDF's page count follows the pictures, not the word count. A 498-character post with four images exported as a 5-page PDF on the same day, with the images placed at full page width. If your PDF looks like it repeats itself, check whether it is image pages rather than duplicated text.

If you are holding a .txt file made before 2026-10-23, it really is truncated: the TXT renderer had no branch for quoted-post notes and dropped them silently, so the 3,672-character thread above exported as 1,212 characters. That is fixed — the same thread now produces 3,456 characters of text — but old files are not retroactively repaired.

3. The download refuses after about three hours, and blames the source

A parsed document is kept for 10,800 seconds — three hours — and then deleted. Nothing is stored longer, so a tab left open over lunch cannot download afterwards. What you see is the confusing part: the request fails with the code doc_expired, but the message shown on the page is “The source took too long to answer.”, which reads like a network problem with X rather than an expired document on this side.

We reproduced it on 2026-10-27 by asking for an export with a document id that no longer exists: the response carried errorCode: doc_expired with errorKey: err_timeout. The fix takes ten seconds — paste the link again so the post is re-read, then export immediately. Waiting, retrying the download button, or changing browser will not help, because the document the button points at is gone.

4. Posts that cannot be read logged out

This tool reads public pages. When a post is not public the export stops with a specific message rather than writing a file with holes in it:

  • Private account — “This account is private, so the content cannot be retrieved.”
  • Deleted, or the id resolves to nothing — “The post was not found. It may have been deleted.”
  • Link resolves but there is no body — “The link resolved, but no article body was returned.” Usually a media-only post or a repost with no text of its own.
  • Too many requests — “You are sending requests too quickly.” Reads are limited to 30 per five minutes and exports to 80 per five minutes; wait a minute rather than re-pasting in a loop.

None of these produce a “partially complete” file, which is worth knowing: if you got a document at all, the missing posts were not silently skipped for permission reasons.

5. A quoted post is a note, not a second thread

When a post in the thread quotes another post, the export writes one block: [quoted @handle] followed by the quoted post's own text. It does not walk into the quoted conversation, so a thread that quotes a 10-post argument arrives with that argument's first post only. In thread A both quoted posts came through in all four formats, as two [quoted @…] notes.

If the quoted post has since been deleted, there is nothing to copy and the note is what remains. That is a limit of the source, not of the export.

A short checklist before you export again

  • Copy the permalink of the last reply in the thread, not the first.
  • Pick the format by what you need next: Word to edit, PDF to send, Markdown to reuse, TXT only when you want characters and nothing else.
  • Export within three hours of reading the link.
  • If a download fails, read the message and match it to the four cases above before retrying.
  • For threads that quote other threads, export the quoted thread separately — it is a different link.

FAQ

Why did my X thread export come out as a single post?

Because the link you pasted was the first post. The chain is rebuilt upward from the link, so only the last reply carries the whole thread. On 2026-10-27 the first post of a 2-post thread returned 891 characters against 3,672 for its last post.

Is the plain-text export missing text?

No. In the 2026-10-27 test the TXT file kept all 12 content blocks of the thread, quoted posts included. It has no images — those become [image: 01.jpg] placeholders. Files exported before 2026-10-23 did drop quoted-post blocks and are genuinely incomplete.

Why does the download say the source took too long?

That message is shown when the parsed document has expired, not when X is slow. Documents are deleted after three hours; paste the link again and export straight away.

Can I export a thread where someone quotes another post?

Yes, and the quoted post's text is included as a [quoted @handle] note. The quoted post's own replies are not followed — export that thread from its own link if you need them.

How long do I have to download the file?

About three hours from the moment the link was read. Re-reading the link creates a new document with a fresh window; the old download button will keep failing.

Find where the missing passage disappeared

  1. Pick one short exact phrase from the first post and another from the last reply you expected.
  2. Look for both in the site preview. If the last one is missing there, verify the reply URL and which posts are publicly visible.
  3. If both appear in the preview, inspect the exported text. In a PDF, also look at the page visually: a text-search failure alone does not prove that the passage is absent.

TweetArticleSave

Paste the last post of an X thread and export it as Word, PDF, Markdown or plain text.

Open TweetArticleSave

More from TweetArticleSave

Save an X thread as a PDF, Word or Markdown (tested) Export an X thread as Markdown Save an X article as a PDF Save an X thread as an image All TweetArticleSave guides