AI Second Brain for Links

About Link Crypt

Link Crypt is the URL layer of an AI second brain — capture from Chrome, let Cryptie auto-tag with AI, then feed the same library to Claude Code, ChatGPT Actions/Codex, Obsidian, and any HTTP agent through the REST API. Built for BASB-style capture and PARA-friendly tag conventions at scale.

New here? Read the AI second brain for bookmarks guide, then explore integrations.

📥
Capture

Chrome extension, iOS share sheet, web app, or POST /links from any agent.

🏷️
Organize

Cryptie reads page context and suggests tags from your existing vocabulary — zero filing required.

🔎
Retrieve

tag:ai, domain:arxiv.org, title:. Filtered search that holds at thousands of links.

🤖
Express

REST API gives Claude Code, ChatGPT Actions/Codex, Obsidian, n8n, and any agent live read/write access. One Bearer token.

Why we built it

Bookmarks scattered across browsers and read-later apps are hard to reason about — especially when you also use Obsidian for notes and AI assistants for research. We wanted one canonical place for URLs: saved with tags, searchable in the app, and addressable over HTTP so an agent can pull context or drop in a new link without copy-paste.

As agentic AI (Claude Code, ChatGPT actions, Cursor) became real in 2025–2026, a well-organized link library stopped being a nice-to-have and became the context layer your agents read. Link Crypt is built for that.

What makes it different

  • Cryptie AI auto-tagging — uses your existing vocabulary, not generic labels.
  • First-class REST API — per-user Bearer keys, OpenAPI spec, and scripted integrations for AI clients.
  • Search built for scaletag:, domain:, combined queries, fast.
  • No ads, no social — your library is private, yours, and fully exportable.
  • Easy migration — import from Chrome/Safari/Firefox/Edge, Pocket, Instapaper, and 10+ others.

Who builds a second brain here

Anyone who accumulates lots of links and wants them on demand — surfaced by their AI when it matters, not buried in folders.

  • Journalists & writers — sources and background reading tagged by story or beat; pull domain: and tag: on deadline.
  • Product & UX — competitors, specs, research sessions; filters stay fast as the library grows.
  • Analysts & researchers — papers, data, dashboards; combined queries like tag:ai domain:arxiv.org stay instant.
  • Developers & PKM builders — Obsidian for notes, Claude Code for reasoning, Link Crypt for URLs over HTTP.

One library — app, API, and agent

The same links you see in the UI are what your Claude project and Obsidian plugin query.

Link Crypt clean dark interface with organized links, AI-suggested tags, and advanced search filters