X / Twitter archive
Request X's official account archive. The owner confirms password and a verification code. There is no supported API replacement for this dump.
Parent argument: data-exports.
Request URL
Start from a signed-in session:
- https://x.com/settings/your_twitter_data
https://x.com/settings/download_your_dataresolves while signed in (verified 2026-09-11).
It re-prompts for the account password, even in a session that is already signed in. This is a deliberate human-presence check, so it cannot be automated, and an agent should hand back to the account owner here rather than handling a credential.
If the direct URL bounces, use the menu. The official label is stable:
More → Settings and privacy → Your account → Download an archive of your data
Submit
Desktop:
- Sign in at https://x.com. Stop. The owner signs in.
- Click More in the left sidebar.
- Click Settings and privacy.
- Click Your account.
- Click Download an archive of your data.
- Enter the account password. Stop. The owner types it.
- Complete the verification code sent to the account email or phone. Stop.
- Click Request archive.
X emails you and shows an in-app notification when the zip is ready. Return to the same Download an archive of your data page and click Download archive. Do this on a desktop. The file can be large.
What it contains
The unzipped archive ships its own inventory. Typical top level:
Your archive.html— offline browser for a subset of the dump. Official archive README (inside the zip) says the HTML renderer only works under 50 GB and does not show everything.data/— machine-readable.jsfiles. Each file is JSON assigned to awindow.YTD.*variable, not bare.json.assets/— icons and fonts for the HTML viewer.
Files the archive README lists as in-scope: profile, posts (tweets), Direct Messages, media attached to posts or DMs, followers, following, lists, inferred interest and demographic data, ads seen or engaged with.
Open data/README.txt (or the equivalent README in that zip) before you trust a file map. X changes the set.
What it does not
- UNVERIFIED: bookmarks. Multiple 2024–2026 archives and tools report that the official zip omits bookmarks. X has not published a current public inventory that confirms this. Do not promise bookmarks. If the owner needs them, say so up front.
- Deleted posts, or media X has already dropped.
- Other people's accounts.
- A live, clickable backup of every attached media URL. Many entries are identifiers or links back to x.com. If X later 404s the media, the archive row still exists and the file may not.
Delivery
X notifies by email and in-app. Typical wait is about 24 hours. Official current help does not publish a hard SLA. UNVERIFIED: waits of 24–48 hours are commonly reported; large accounts take longer.
You download from the same settings page, not from a raw attachment.
UNVERIFIED: the download window is about a week. If the Download button has vanished, request a new archive.
Format
One ZIP. Inside: HTML viewer plus data/*.js.
Notable data/ files, when present:
| File | Holds |
|---|---|
tweets.js or tweet.js |
Posts, replies, reposts |
like.js |
Likes |
direct-messages.js |
DMs |
follower.js / following.js |
Graph |
account.js / profile.js |
Account and profile fields |
Strip the window.YTD.… = prefix before you parse a file as JSON.
Gotchas
- Password plus a one-time code. The agent stops at both.
- Download from the settings page, not only from the email. The mail is a heads-up.
- HTML viewer is incomplete. Always open
data/. .jsnot.json. Parsers that expect raw JSON will fail until you strip the assignment.- No polling. Watch the inbox. X does not expose a ready-check API for this.
- EU path. Some third-party write-ups say EU accounts must make a separate request. UNVERIFIED against current official help. If the settings page refuses, stop and read what it says.
Sources
- In-product path: Settings and privacy → Your account → Download an archive of your data
- Archive README shipped inside the zip (authoritative for that request)