FileOctopus Find words inside documents
- Open FileOctopus → Settings… and add the folder to Search if needed.
- Turn on Contents for that folder.
- Allow indexing to complete, then use a query such as
content:"launch checklist".
The local extractor supports common plain-text and source formats, including TXT, Markdown, CSV, JSON, logs, and code, plus the text layer of PDFs. It is not a general extractor for every Office or proprietary format. Unsupported files can remain filename-searchable.
Local extraction is bounded: plain text reads up to 2 MiB and stores up to 200,000 characters; PDF text extraction considers up to 50 pages, skips PDFs above 64 MiB, and has a processing budget. A match beyond those limits may be absent.
Find text in images and scanned PDFs
Enable Text in Images separately for the folder. OCR uses Apple's on-device Vision recognition. Use content: to search the recognized text; this is text recognition, not image-subject or semantic search.
Recognition depends on image quality and language support on the running OS. The implementation prefers English and supported Vietnamese recognition languages. OCR is bounded to eligible local images and up to 10 scanned-PDF pages, with a 50 MiB file limit, image dimension limits, and a 200,000-character storage limit.
Locked, corrupt, unsupported, unreadable, or nonlocal cloud files can be skipped. The app does not intentionally download placeholders to build its content/OCR index.
Spotlight and local indexes
Find can combine filename results with system Spotlight results and the app's opt-in content/OCR providers. Spotlight availability and metadata depend on macOS indexing. FileOctopus's per-folder Contents and Text in Images controls govern its own extraction; they do not change macOS Spotlight settings.
Read result/provider status when coverage is incomplete. There is no promise that an unindexed document's contents will be found, even when its name appears.
Find files by recorded source
Try downloaded from github, from:github, or another supported source token. FileOctopus classifies recorded origin information such as where-from URLs and quarantine metadata; the source-details inspector explains available evidence.
A label such as GitHub, Gmail, or Google Drive is not an account integration. FileOctopus does not log into these services or search remote messages and repositories. A copied file can lose its origin metadata, and not every downloaded file carries it. A missing source match does not prove the file came from somewhere else.
Source filters do not imply a reliable download timestamp. Use explicit creation or modification filters for date constraints, and inspect their chips.
Control retained content
Turn off a folder's content/OCR option or remove the root when it should no longer contribute that indexed content. Erase Index… removes FileOctopus's stored search data; retained approved folders will be indexed again. See index settings and privacy.
Watch: turn a phrase into filters
A real Find session: combine kind, date and size, broaden the query by removing a chip, then try recorded origin and a supported Vietnamese phrase.
Recorded in FileOctopus using fictional files. No audio. Playback is 1.25×. GitHub origins are synthetic metadata in the demo files; no account is connected. Result counts depend on the fixture and recording date.
Read the demo steps
- In Find, use This Mac, which is restricted to the fixture's approved folder in this recording.
- Enter ‘pdfs from last week over 5 mb’. The app creates PDF, Last week and > 5 MB chips. This fixture has no matching files.
- Open the size chip menu and remove that condition. ‘pdfs from last week’ returns seven matching fixture files.
- Enter ‘downloaded from github’. Two files with recorded GitHub origin appear.
- Enter ‘tệp PDF tuần trước’. The PDF and Last week chips return the same seven files as the English date-and-kind query.