You download your X data archive (Settings → Your Account → Download an archive), upload the ZIP, and X Brain processes everything:
– Semantic search – search by meaning, not keywords. Type “advice on dealing with imposter syndrome” and it finds the right tweets even if none of them use those words.
– Auto-classification – 15 knowledge domains, 65+ subcategories. Engineering, career advice, marketing, finance, and so on. Runs automatically, no tagging required.
– Knowledge extraction – AI pulls out key takeaways, tools mentioned, and references from each tweet.
– Obsidian export – downloads your whole library as structured Markdown with frontmatter, tags, and index files.