A self-contained, offline-friendly search page for your Twitter/X archive. No server, no database, no dependencies — just a Python build script and a single HTML file you open in any browser.

• Full-text keyword search across all your tweets
• Filter by year, "has photo", "has video", "replies only", or "original only"
• Inline photo grid and video player on each tweet card
• Lightbox for full-size image viewing
• Sort newest or oldest first
• Works entirely offline — no data ever leaves your computer

Go to Settings → Your account → Download an archive of your data on Twitter/X. You'll get an email with a .zip file when it's ready. Unzip it.

At minimum you need your tweet data files. Copy whichever of these exist in your archive:

For photos and videos, copy these folders too (optional but recommended):

Your folder should look like:

tweet-search/
├── build.py
├── tweet.js          ← your archive data
├── tweets_media/     ← your photos (optional)
└── archive_tweet_media/  ← older photos (optional)

python3 build.py

This reads all your .js files, builds an in-memory index, maps photos/videos to tweets, and writes a single tweet_search.html file. It takes about 10–30 seconds depending on archive size.

Open tweet_search.html in any browser. Everything is embedded — you can move the HTML file anywhere as long as the tweets_media/ and/or archive_tweet_media/ folders stay next to it.

At the top of tweet_search.html (after running the builder), find this block and update your name and handle:

<div class="display-name">Your Name</div>
<div class="screen-name">@yourhandle</div>

Or edit build.py and change the two placeholder strings in build_html() before running.

• Tweet data: All tweet JS files use one of three formats (newer window.YTD.tweets, flat window.YTD.tweet, or old Grailbird monthly). The builder handles all of them, deduplicating by tweet ID.
• Media mapping:
  • tweets_media/ files are named TWEETID-MEDIAID.ext, so the link is direct.
  • archive_tweet_media/ files are named with a SHA-256 hash of the original https://pbs.twimg.com/media/… URL. The builder re-derives that hash from each tweet's media entities to find the right file.
• Search: All tweet text is embedded as JSON in the HTML. Search is pure client-side JavaScript — instant, no server needed.

• Python 3.7+ (standard library only, no pip install needed)
• A Twitter/X data export

MIT — do whatever you like with it.