07: Read your own X archive

Time: 15 minutes You need: your X data export. No login, no API key, and nothing leaves your machine. You end up with: your whole account as JSON, CSV, Markdown and a browsable HTML page, and a staged migration to Bluesky or Mastodon.

Everything else in these tutorials reads X over the network. This one reads the file X gives you, which is the one copy of your account that nobody can rate limit, suspend, or take away.

1. Ask X for the export

X does not hand it over instantly. Request it first, then come back:

1. x.com → Settings and privacy → Your account → Download an archive of your data
2. Confirm your password, then confirm the code X sends you.
3. Wait. X takes anywhere from a few hours to a couple of days.
4. Download the zip when the notification arrives. It is called something like twitter-2026-08-28-a1b2c3....zip.

Do not unzip it. XActions reads the zip directly, and an unzipped archive is just as good if you already opened it.

2. See what is in there

xactions archive summary twitter-2026-08-28.zip

✔ Read twitter-2026-08-28.zip (zip)

  X archive for @yourhandle (zip)
  Account created: 2019-03-01
  Tweets span:     2024-01-01 to 2025-03-15

  Tweets       3 (3 original, 0 replies, 0 retweets, 0 with media)
  Likes        1
  Following    2
  Followers    1
  Blocks       0
  Mutes        0
  DMs          0 messages in 0 conversations
  Lists        0
  Media files  0
  Engagement   12 likes, 3 retweets received
  Busiest year 2024 (2 tweets)

  Tweets per year
    2024  2
    2025  1

  Top hashtags
    #xactions  2
    #firstpost  1

  Top mentions
    @alice  1

  Not in this archive: blocks, mutes, dms, lists, media

  Export it with `xactions archive export twitter-2026-08-28.zip --out <dir>`.

(That run is against a three-post test archive, so the numbers are small. Yours will not be.)

The Not in this archive line matters. X's export contents vary by account age and by what you have used, so a missing section is normal rather than a parse failure. archive summary tells you which sections it actually found instead of quietly reporting zero.

Two flags worth knowing:

xactions archive summary twitter-2026-08-28.zip --top 25          # more hashtags and mentions
xactions archive summary twitter-2026-08-28.zip --sections tweets # skip the rest, much faster

--sections is the one to reach for on a large archive: reading only tweets skips megabytes of DMs and likes you did not ask about. It also means the account header reads @unknown, because the handle lives in the account section you just skipped.

And as data, for anything that is going to script this:

xactions archive summary twitter-2026-08-28.zip --json | jq '.counts'

{
  "tweets": 3,
  "replies": 0,
  "retweets": 0,
  "original": 3,
  "withMedia": 0,
  "likes": 1,
  "following": 2,
  "followers": 1,
  "blocks": 0,
  "mutes": 0,
  "dmConversations": 0,
  "dmMessages": 0,
  "lists": 0,
  "mediaFiles": 0
}

The other top-level keys are username, accountCreatedAt, dateRange, busiestYear, tweetsPerYear, engagement, topHashtags, topMentions, sections, format, archivePath, and source.

3. Turn it into files you can actually use

X's export is a folder of JavaScript files that assign to window.YTD.*. That is fine for X's own viewer and useless for everything else. archive export converts the lot:

xactions archive export twitter-2026-08-28.zip --out my-archive

✔ Read twitter-2026-08-28.zip (zip)
✔ Wrote 25 files to my-archive

  profile      1
  tweets       3
  likes        1
  following    2
  followers    1
  blocks       0
  mutes        0
  dms          0
  lists        0
  media        0

  Open my-archive/index.html in a browser to browse it.
  Compare against a live export with `xactions diff <dirA> my-archive`.

You get four shapes of the same data, because different jobs want different ones:

--formats narrows that when you only want one:

xactions archive export twitter-2026-08-28.zip --out my-archive --formats json,md

--sections works here too, so --sections tweets,likes gets you a posts-only export in a fraction of the time.

Now the useful part. This is ordinary JSON:

jq 'length' my-archive/tweets.json
jq -r 'sort_by(-(.likes // 0))[:10][] | "\(.likes)\t\(.text[0:80])"' my-archive/tweets.json
jq -r '[.[] | select(.text | test("hiring"; "i"))] | length' my-archive/tweets.json

4. Diff two exports

Request a second archive six months later and the two are directly comparable:

xactions diff my-archive-january my-archive-july

That is also how you check what a live scrape missed: export your account with xactions export YOUR_HANDLE, then diff that directory against the archive export. Anything present in the archive and absent from the scrape is content X no longer serves publicly.

5. Take it somewhere else

The archive already holds everything needed to rebuild the account elsewhere, so migration is a read of a local file rather than thousands of scraped requests.

Dry run first. It is the default, and it writes nothing to any network:

xactions archive migrate twitter-2026-08-28.zip --to bluesky --out staged

✔ Migration preview complete

  Platform: bluesky
  Mode:     DRY RUN
  Tweets:   3/3 ready
  Follows:  2/2 matchable
  Staged:   staged

  Sample actions:
    create_post: Hello world #firstpost [dry-run]
    create_post: Shipping #xactions with @alice [dry-run]
    create_post: Still #xactions [dry-run]
    follow: unknown [dry-run]
    follow: unknown [dry-run]

  This was a dry run. Add --execute and your credentials to perform it.

--out writes the normalised archive into the staged directory: tweets.json, following.json, followers.json, likes.json, profile.json, and a summary.json, one file per section. Read them. Edit tweets.json. This is the moment to drop the 2019 posts you would rather not carry over, and it is much easier than deleting them on the far side. Each staged post carries id, text, createdAt, url, inReplyTo, retweeted, media, and metrics.

Mastodon is the same command with a different target, and posts become toots:

xactions archive migrate twitter-2026-08-28.zip --to mastodon --out staged

    create_toot: Hello world #firstpost [dry-run]

Then, and only then, execute:

# Bluesky: use an app password, never your account password.
# Settings → Privacy and security → App passwords.
xactions archive migrate twitter-2026-08-28.zip \
  --to bluesky --execute \
  --handle you.bsky.social \
  --password xxxx-xxxx-xxxx-xxxx

# Mastodon: a token from Preferences → Development → New application,
# with write:statuses and write:follows.
xactions archive migrate twitter-2026-08-28.zip \
  --to mastodon --execute \
  --instance https://mastodon.social \
  --token YOUR_ACCESS_TOKEN

--json on any of the three sub-commands prints the summary as data, which is what you want if this is running unattended.

6. Where the archive beats scraping

Worth being clear about, because it decides which tool to reach for:

The follower row is the one that surprises people. X's export stores your followers as account ids, not handles, so the archive tells you how many and which ids, and a live read is what turns those into names.

What you learned

• The archive is the only copy of your account nobody can rate limit or revoke
• archive summary tells you which sections X actually included, not just counts
• archive export turns window.YTD.* JavaScript into JSON, CSV, Markdown and a browsable page
• --sections and --formats are how a large archive stays quick
• Migration is a dry run by default, and staging the actions is your chance to edit them

Next

• 06: Everything is JSON to script the exported files
• 05: Read any account like an analyst for the live-data equivalent
• Account portability for the migration internals
• CLI reference for every flag