xmark vs native X bookmarks
X gives you a list. xmark gives you a library.
why xmark?
Semantic search, not just keyword
X's search wants exact words. xmark uses OpenAI embeddings so you can ask in any phrasing and the right tweet surfaces.
Chat with citations
Ask the librarian a question — 'what did I save about agents last month?' — and get an answer with inline citations back to the original tweets. X has no chat layer.
Full-content indexing
xmark indexes the full tweet text including long-form notes, quoted tweets, image alt-text, and linked article metadata. X search treats most of that as opaque.
Developer API + MCP
REST API and MCP server let your own agents query your bookmark library. X's bookmark surface is read-only and not exposed to third parties.
feature comparison
| feature | xmark | native X bookmarks |
|---|---|---|
| save bookmarks (read access) | via OAuth | native |
| reverse-chronological list | ||
| keyword search | limited | |
| semantic search (embeddings) | ||
| AI chat with citations | ||
| filter by date / topic | limited | |
| developer REST API | ||
| MCP server for AI agents | ||
| export to JSON / Markdown | ||
| price | $5/month | free with X |
the verdict
X's native bookmarks are fine if you have a few dozen saves and never need to find anything. xmark is for when the pile got too big to scroll and you want to actually use what you saved. The bookmarks stay on X — xmark is a search-and-chat layer on top.
ready to get started?
sign up in seconds. free to start, no credit card required.