Search
Open the app: Search inside a PDF → (products.documentize.app)
In one sentence
Search looks for a word, phrase or pattern inside your PDF and gives you the PDF back with every match highlighted and a list of results you can jump through.
What makes it unique
- You do not get a list of hits on a web page: you get your own PDF back, marked up. Every hit is highlighted, each paragraph containing a hit gets a frame, and a "Search results" bookmark tree is added so you can jump from paragraph to paragraph in any PDF viewer.
- It can search with regular expressions (patterns such as "any number followed by EUR").
- It can search scanned PDFs by running recognition first (tick one box and choose the language).
- The search is not case-sensitive.
How it differs from its look-alikes:
- Search (this app) searches inside the PDF files you upload and returns marked-up PDFs.
- Global Search is something else: it is the search box of the Documentize site itself (it finds tools and pages on the site), not a way to search your documents.
- Extract Text gives you all the words but does not find anything.
When to use it / when not to
- Use it to locate every mention of a client, a clause, an amount or a keyword in a long PDF, and to hand a colleague a copy with the hits already marked.
- Do not use it if you just want to read the text elsewhere (Extract Text), or to ask questions in natural language (Ask PDF).
- Search matches the words as written; it does not understand synonyms.
Before you start
- A PDF with a text layer, or know that it is a scan (then tick the scanned box).
- For scans, know the document language.
- Decide whether you want a plain search or a pattern.
Step by step
- Open Search ("Search in PDF") and upload your PDF(s).
- Type your search text in the box. (If you press the button with the box empty, the page asks you to "enter text or a regular expression".)
- Optionally tick Find by regular expression if the text is a pattern.
- If the PDF is a scan, tick My PDF documents are scanned and choose Documents language.
- Press Search.
- Download the marked PDF; open it and use the bookmarks panel. Search again starts over.
📷 Screenshot placeholder: the search box with the two checkboxes
Options explained
| Option | What it does | Default | Recommended |
|---|---|---|---|
| Search text | The word, phrase or pattern to find | empty | Short and specific; special characters like + or ( are matched literally when regular expressions are off |
| Find by regular expression | Treats your text as a pattern instead of plain words | off | On only if you know the pattern; to match several different words, the site's FAQ suggests a pattern that lists them |
| My PDF documents are scanned | Runs text recognition first so a scan can be searched | off | On for scans and photos saved as PDF |
| Documents language | The language used by the recognition (shown with the scanned option) | English | Always match the document |
What you get
- A PDF with highlighted matches, frames around the paragraphs that contain them, and a Search results bookmark tree.
- With several PDFs, only those that contain a match are returned (as a ZIP when more than one).
- If nothing is found anywhere, there is no file; the page tells you that no results were found.
- The original is not modified; you receive a marked copy.
Quality & limits
- The marked copy is the same document with annotations added: content is not recognized or rewritten, except that for scanned files the recognition step runs first, and the same accuracy limits as Searchable apply.
- On a scan, a word the recognition got wrong will not be found, so "no results" does not prove absence.
- Words split across lines or hyphenated may be missed.
- Free-plan time limits and file-count/size limits apply as for other apps (see the site).
- Results are kept for about two days.
Works well with
- Use before: Searchable (do it once, then search many times without waiting for recognition), Unlock, Rotation.
- Use after: Splitter to cut out the pages with hits, Compressor to shrink the marked copy.
- Recipes: Digitise an archive, Share an excerpt.
Common mistakes & fixes
- "No results, but I can see the word" → the PDF is a scan (tick the scanned box and set the language), or the word is part of a picture, or the spelling differs.
- "A symbol makes my search fail" → with the regex box ticked, symbols are pattern commands; untick it for plain text.
- "I searched five files and got one back" → only files with hits are returned.
- "I expected a list of pages" → look in the bookmarks panel of the returned PDF.
Same thing elsewhere
- CLI plugin: Documentize.Search (query and optional regular expression).
- MCP tool: SearchInDocument.
- See Beyond the browser.
Try it
Take a text PDF of at least ten pages. Search for a word that appears several times. Open the result, look at the highlights and open the bookmarks panel to jump between hits. Then search for a pattern such as a digit sequence with the regular expression box ticked and compare.