Full-Text Search
As your personal knowledge base grows, so does the information within it. DoCube's powerful and intelligent search is designed to help you instantly find what you need across a vast collection of notes, books, web pages, and documents, whether you recall exact words or just a vague concept. The full-text search feature in DoCube offers the following capabilities:
- Supports Multiple File Formats: Including MinerU, Markdown, PDF, EPUB, and HTML.
- Global Search: Search across all documents and notes in your knowledge base, with results ranked by relevance.
- Multi-dimensional Filtering: Quickly narrow down results by tags, document name, document type, or text source.
- Real-time Preview: Get a real-time preview with keyword highlighting, allowing you to quickly skim content and decide whether to open the full document.
Building the Index
To enable efficient full-text search, DoCube first needs to index imported documents. Indexing happens automatically in the background without any manual intervention. When document content changes, DoCube also periodically re-indexes the modified documents. The status icon 
to the right of each document title on the Documents page indicates its indexing status: gray means not yet indexed, blue means indexing is complete. If an expected document does not appear in search results, you can check its indexing status and try the search again after indexing is finished.
Initiating a Full-Text Search
Starting a search in DoCube is quick and easy. Simply click on the title bar of any tab, such as the Home tab title bar shown below, to immediately bring up the search box and begin a new search.
Search in DoCube
After completing your search, you can always click the 
icon in the navigation bar to seamlessly return to the document you were just reading.
Filtering Search Results
After the initial search, you can click the 
icon in the top toolbar of the search results page to open the filter panel. The various filter options are explained below:
Expanded query
After the initial query, DoCube extracts some related or synonymous expansion terms from the found content. You can click one or more of these terms to combine them with your original query, forming a new, richer query to retrieve more relevant documents from your collection.
Filter by Term
Show only results that must contain the selected term(s). Multiple terms can be selected for filtering.
Filter by Content Source
Show only results where the search term appears in the specified source(s). For example, show only results where the term appears in the title. Multiple sources can be selected.
Filter by format
Show only results from the specified document type(s). For example, show only results from PDF documents. Multiple types can be selected.
Filter by document
Show only results from document(s) with the specified name(s). Multiple documents can be selected.
Filter by tag
Show only results that contain the specified tag(s). The resulting documents must have the selected tag(s). Multiple tags can be selected.
Quickly Obtaining Precise Search Results
Facing a large document library, a single search might return many results. DoCube displays only the top 40 results most relevant to your search query to help you quickly sift through content. The top-right corner of the search results page shows statistics for the current search, including the total number of results and the number of relevant results, as shown below.
If your expected content isn't among the first 40 results, you can use the filters described in the previous section to further narrow the search scope. When the total number of results drops below 500, the total count in the statistics will turn blue and become underlined. You can then click this number to switch to viewing all search results, navigating through them using the pagination buttons at the bottom of the search page.
Of course, looking through nearly 500 results is still not very efficient. A more effective approach is to continue using filters to narrow the scope. When the number of relevant results matches the total number of results, it indicates the current search results are highly precise. If your expected content is still not found, it likely means the knowledge base does not contain relevant information.