A Next.js 16 dashboard for browsing, searching, and tracking liked tweets from your Twitter data archive exports.

Twitter's Data Privacy feature lets you download your complete account data as a self-contained archive. This project provides an interactive UI to explore those archives without being logged in for every action.

• Multi-archive support — load multiple Twitter data exports and browse them individually or as a single deduplicated view
• Full-text search — filter tweets by keyword in real time
• 4-state read tracking — cycle each tweet through Read → In-Progress → Ignore → Unread; batch save pending changes
• Bookmarks — star any tweet to bookmark it; dedicated Bookmarks tab filters to bookmarked tweets only
• Statistics — read/unread progress bars, overall and broken down by year
• Word cloud — most frequent words across all tweet text; click a word to see matching tweets; dismiss irrelevant words
• Year navigation — year divider headings in the tweet list with a sticky side panel for quick jumps on wide screens
• Markdown export — download any archive view as a .md file
• URL expansion — t.co short links are expanded to their real destinations
• Preflight checks — automatic dependency and build-cache validation before npm run dev

• Node.js v18+
• npm v8+
• One or more Twitter data archives (exported via Twitter's data download)

Place your exported Twitter archive folders inside zip/:

zip/
├── twitter-2022-02/
├── twitter-2023-12/
├── twitter-2024-11/
└── ...

Ensure each folder contains data/like.js. Archives are detected automatically.

cd dashboard
npm install

Expands t.co short URLs to their real destinations:

npm run expand-urls

This is resumable — safe to interrupt and re-run. Dead links are stored as null and shown as [Dead link] in the UI. Previously dead links are automatically rechecked on each run.

npm run dev

Open http://localhost:3000 in your browser. A preflight check runs automatically before the dev server starts, validating dependencies and build cache.

Produces a single Markdown file of all liked tweets with read-state and bookmark fields for tracking:

npm run generate-consolidated

Output: asset/CONSOLIDATED.md. Append-only — existing entries and manual edits are preserved on re-run.

Precomputes the top 500 most frequent words for the Insights tab:

npm run generate-word-cloud

Output: asset/word-cloud.json and asset/IGNOREDWORDS.md. User-dismissed words are preserved across regenerations.

twitter-archive-explore/
├── zip/                        # Twitter data archives (read-only)
├── dashboard/                  # Next.js 16 app (React 19, Tailwind 4)
│   └── src/
│       ├── app/                # Pages and API routes
│       ├── components/         # React components
│       ├── lib/                # Parsers, utilities, word cloud logic
│       ├── scripts/            # CLI scripts (expand-urls, generate-*, preflight)
│       └── types/              # TypeScript type definitions
├── asset/                      # Generated files (gitignored)
│   ├── url-cache.json
│   ├── word-cloud.json
│   ├── IGNOREDWORDS.md
│   └── CONSOLIDATED.md
├── images/                     # Screenshots for README
└── scripts/                    # Git hooks (pre-commit)

All commands run from inside dashboard/:

For detailed technical reference, see CLAUDE.md. For step-by-step setup instructions, see INSTRUCTIONS.md.

MIT, (c) Sourav