Which folders should you add to local document search?

Start a local document index with the folders that contain the work you actually need to find. A small, relevant collection is easier to check than an entire drive full of caches, duplicates and unrelated files.

The aim is a useful search scope, not the largest possible index.

Begin with one active project

Choose a folder containing a brief, notes and a few documents you recognize. Avoid mixing personal records into the first trial. Make sure cloud-synced files are actually downloaded and readable on the computer.

In CiteSilo, add that folder to local file search and let indexing finish. Search for a phrase you know is inside one of the files, then open the result. This checks the path from selected folder to extracted content to source access.

If the known file is missing, investigate that before adding more folders.

Know which files are eligible

CiteSilo 0.1.7 indexes supported text formats, source-code files, PDF and DOCX. Its current filter excludes empty files and files larger than 4 MiB. It also skips many build, dependency, cache and hidden folders.

An index is therefore not a mirror of every file beneath the selected directory. A spreadsheet workbook, image collection or large PDF may need a different search approach. Do not rename an unsupported file’s extension to force it through an extractor.

Some readable configuration formats are eligible too. Keep credentials and unrelated configuration outside the selected scope rather than assuming a document index automatically recognizes secrets.

Organize by retrieval need

A useful starting structure is an active-project folder, a reference folder and a small archive of completed work. Keep the scope understandable enough that you can explain why a result is present.

If several files contain the same text, inspect the full location and revision before choosing one. A matching passage in an obsolete copy is still a match, but it may be the wrong source for today’s task.

Folder organization does not have to be perfect before search becomes useful. Focus on separating material with different purposes or access requirements.

Verify extraction with real examples

Test one representative file of each type you rely on. A PDF with selectable text and an image-only scan behave differently. A DOCX with a paragraph of prose is different from a document dominated by embedded images.

Use the result excerpt to locate the passage, then read it in the original file. Search quality should be judged on your actual documents rather than a generic claim about supported formats.

Expand only when the first scope helps

Add another folder when you can name a recurring retrieval problem it will solve. Revisit the scope when a project ends or confidentiality requirements change.

For a practical example, read how to find a PDF without its filename. The local-first guide explains where the index and optional AI processing fit.