Linkly AILinkly AI
Back to blog

v0.3.0 Release: Library Management, Finally Complete

·Linkly AI·
Releasev0.3.0Library ManagementCloud SyncSemantic Search

Today we’re releasing another milestone — Linkly AI v0.3.0, a major upgrade focused on library management.

Linkly AI’s core value is helping you efficiently manage and use your local files. That’s why, early on, we didn’t rush to build complex library management. Instead, we focused on building a powerful, stable, low-cost indexing and search engine, along with a small but carefully polished set of AI tools.

Not having library management actually led to an unexpected benefit. We found that many users simply threw all their local documents, sometimes tens of thousands of them, straight into Linkly AI. With global search and tools like outline, this actually simplified traditional knowledge management workflows significantly.

After nearly a month of beta testing, from the feedback we’ve received, many professional users have already integrated Linkly AI into their daily workflows. Even more exciting, some users have built integrations with Zotero, the professional reference management software, so they can include their Zotero libraries in Linkly AI’s search scope.

But as document collections grow, users start facing new challenges:

  • With too many documents, search accuracy and efficiency suffer
  • Professional users doing focused research find their work documents mixed in with personal files, which can cause unnecessary headaches
  • Some users have to keep their computers running 24/7 just to ensure their document library is always searchable

So in v0.3.0, we’re introducing library management, with cloud sync for libraries coming gradually in future iterations.

Here’s a quick overview of what’s new in v0.3.0.

Library Management

This is the core feature of v0.3.0.

In the Settings panel, you can now create, edit, and delete libraries. Each library is simply a collection of folders — you can organize them by project, topic, or whatever makes sense to you.

For example, if you’re a researcher, you could put your papers in an “Academic Literature” library, work documents in a “Company Projects” library, and personal notes in “Daily Notes.” When searching, work stays with work and personal stays with personal — no cross-contamination.

For users who’ve already thrown tens of thousands of documents into Linkly AI, library management can significantly improve search precision — narrow the search scope, and you’ll naturally find what you need faster. And if you’ve already added directories through the folder menu, adding them again in a library won’t cause duplicate indexing.

With libraries in place, search tools need to keep up.

Type @ in the search launcher and a library list will pop up. Select a library, and your search is scoped to that library. Want to exit? Just press Backspace.

We also added a document type filter panel — when search results contain multiple document types, the panel appears automatically. You can click to filter or use Cmd/Ctrl+↑/↓ to quickly switch between types.

Search Experience Improvements

Beyond library search, we’ve polished the search experience in several ways:

  • Keyword highlighting: Matched keywords in titles and paths are now highlighted
  • Smart path shortening: Long file paths are automatically shortened so they don’t stretch the UI
  • Result limit increase: Bumped from 20 to 100 results — users with large collections won’t see their results cut off
  • Search result caching: Repeated searches for the same keyword are cached for 10 seconds, so you don’t have to wait again
  • Reveal file on open: Clicking a search result to open a file will also reveal it in your file manager

Agent Tool Enhancements

Linkly AI’s MCP tools and CLI are important daily tools for many professional users, and they’ve been upgraded alongside the library feature.

New MCP tools:

  • list_libraries: Lets AI Agents list all your libraries
  • explore: A compact but powerful overview tool — AI Agents can use it to understand topic distribution, directory structure, and top keywords in your library, perfect for getting the big picture or doing survey-style research
  • search now supports library and path_glob filter parameters for searching within a specific library or path pattern — when an Agent discovers a particularly useful folder, it can focus its search there

New CLI commands:

  • linkly list-libraries: List all libraries
  • linkly explore: View library overview
  • linkly search --library "Academic Literature": Search within a specific library
  • linkly doctor: A self-diagnosis command — run it when you hit connection issues and it’ll tell you what’s wrong. Especially handy for Agents like OpenClaw to troubleshoot on their own

Other Improvements

  • Settings panel language detection is more accurate — no more English UI on a Chinese system
  • Fixed an edge case in DOCX document heading extraction
  • Improved keyword extraction quality by filtering out URLs and file paths
  • Launcher footer shortcut hints can now be collapsed, giving more room for search results
  • Fixed remote tunnel connection timeouts, custom port issues, and a CLI tunnel hang issue

Full changelog is available here.

If you haven’t tried Linkly AI yet, you can download it for free from the official website — available on macOS, Windows, and Linux.