Best tools to search your X likes in 2026
Short answer: for your whole like history searchable by meaning, use a tool that reads the official X archive, such as X Brain ($4.99 once). For saves captured automatically as you go, use a syncing extension such as Dewey or Twillot. For bookmarks inside a reading workflow, use Readwise Reader. The catch with extensions is that they only see what X’s likes page still loads, which stops short of the oldest likes on large accounts.
We make X Brain, so weigh our pick accordingly. Every price below was checked on the vendor’s own page on the date shown in its row.
Which tool for which job
X Brain: your whole like history, searchable by meaning
It reads like.js from the official X archive, which holds every like back to the first, so it is not limited by how far X's likes page will scroll. Search works by meaning, and the library exports to an Obsidian vault and connects to Claude.
Dewey: ongoing capture across X and other platforms
It keeps syncing new saves and covers eight more platforms, and because it reads the page it keeps each post's author, which the X archive does not store. Old likes depend on how far X's likes page still loads.
X Brain vs Dewey, in detailReadwise Reader: bookmarks inside a reading and highlighting workflow
It syncs X bookmarks into a read-later app with strong highlight export to Obsidian and Notion. It has no import for likes, so it suits people who bookmark rather than like.
X Brain vs Readwise Reader, in detailTwillot: a cheap local extension for bookmarks, likes and your posts
It syncs bookmarks, likes and your own posts in one Chromium extension, from $2.49 a month, and exports Markdown. Each plan caps the number of posts, and search is full-text rather than by meaning.
X Brain vs Twillot, in detailContextBolt: bookmarks with a Claude connection on a monthly plan
It auto-captures X bookmarks and its Pro plan includes a personal MCP endpoint for Claude, Cursor and Codex. Its pricing page does not mention likes.
Twibird: tagging likes and bookmarks in a free extension
A Chrome extension that syncs likes and bookmarks with tags and saved filters. Paid plans raise the sync cap, which is still bounded by what X's web app will load.
X Brain vs Twibird, in detailFabric: a general AI workspace that also holds tweets
Tweets come in one at a time as you clip them, so your past likes are not imported, but it has AI search across everything you store and an official MCP server.
X Brain vs Fabric, in detailSide by side
| Tool | Price | Old likes | How it gets your posts | Search | Export | Claude (MCP) | Checked |
|---|---|---|---|---|---|---|---|
| X Brain | $4.99 once | Yes, the full history from the X archive | Upload the X archive ZIP; re-upload to refresh | Semantic search plus AI categories | Obsidian vault, Markdown, CSV, JSON | Yes, MCP server | 2026-10-09 |
| Dewey | Free (manual, 1 account); $10/month or $7.50/month billed annually; $225 lifetime | Yes, via its browser extension, as far back as X's likes page loads | Browser extension, continuous sync on paid plans | Keyword search and filters | PDF, CSV, media, Notion, Google Sheets | Not listed | 2026-10-09 |
| Readwise Reader | $9.99/month billed annually ($12.99 monthly); Lite $5.59/month billed annually | No, bookmarks only | Account integration and browser extension | Search within Reader | Obsidian, Notion and more | Not listed | 2026-10-09 |
| Twillot | Free (100 posts); $2.49, $4.99 or $9.99/month | Yes, up to your plan's post cap | Chromium browser extension | Full-text search, filters, smart folders | Markdown, PDF | Not listed | 2026-10-09 |
| ContextBolt | Free (150 bookmarks); Pro $6/month | Not listed (bookmarks) | Auto-capture and in-feed save | Instant on-device search | CSV, JSON, XLSX (Pro) | Yes, MCP endpoint on Pro | 2026-10-09 |
| Twibird | Free up to 1,000 items; from $8.90/month | Yes, up to the plan cap and what X's web app loads | Chrome extension | Keywords, tags, filters | CSV, Notion | Not listed | 2026-09-13 |
| Fabric | From $8/month, billed yearly | No, individual clips only | Clipper and app | AI search across your workspace | HTML | Yes, official MCP server | 2026-09-13 |
Why old likes are the hard part
X’s likes page loads in pages and, on accounts with large histories, stops returning older entries past a certain depth. Any tool that reads that page inherits the limit. The data archive X lets you download has no such limit: its like.js file lists every like with the post’s ID and text, though not the author or the date you liked it. You can open it yourself with the free X Archive Viewer or read what like.js contains.
See also: every X Brain comparison.