Extract, Read & Search
The idea: a PDF is often a locked box. Inside there may be words, pictures, form answers, comments, attached files, or only a photograph of a page. Apps in this family get the content out or make it findable. They produce text, images, spreadsheets, a marked-up copy or a report. They never change your original.
Two kinds of work live here:
- Make it readable. A scan is only a picture. OCR Image and Searchable turn pictures of text into text.
- Take it out or find it. The Extract apps, Parser and Search work on text that is already there.
The apps
| App | Takes | Gives | One-liner |
|---|---|---|---|
| OCR Image | images (JPG, PNG) | .txt files | Read text from photos and scans |
| Searchable | scanned PDF | PDF (or Word, Excel, PowerPoint) | Same pages plus an invisible text layer |
| Extract Text | PDF with text | .txt | The words already in a PDF, three formatting modes |
| Extract Image | PNG images | Every embedded picture | |
| Extract Metadata | on-screen table + CSV | Author, dates, producer, custom properties | |
| Extract Form Data | filled PDF form | CSV files | Field names and answers |
| Parser | ZIP | Text, images, attachments, annotations, bookmarks, form data at once | |
| Search | PDF(s) | PDF with highlighted hits | Find words or patterns, scanned PDFs too |
| Global Search | a few typed words | list of site pages | Finds tools on the Documentize site (does not search your files) |
| Resume Analyzer | CV (PDF/DOC) | PDF report | AI check of a CV against a job (AI site) |
Choosing quickly
- I have photos of text and want the words → OCR Image
- I have a scanned PDF and want to search or copy from it → Searchable
- I have a normal PDF and want the text → Extract Text
- I want the pictures in a PDF → Extract Image
- Who wrote this file, and when? → Extract Metadata
- Answers from returned forms → Extract Form Data
- Everything inside, once → Parser
- Where does this word occur in the PDF? → Search
- Which Documentize tool does X? → Global Search
- Is this CV a fit for that job? → Resume Analyzer
The look-alikes in one table
| Input | Output | Text layer? | Reads pixels? | |
|---|---|---|---|---|
| OCR Image | pictures | .txt | no PDF produced | yes |
| Searchable | scanned PDF | PDF with hidden text | adds one | yes |
| Extract Text | PDF with text | .txt | uses the existing one | no |
| Parser | ZIP of many parts | uses the existing one | no | |
| Extract Form Data | form PDF | CSV | not relevant | no |
Rules of thumb
- If you cannot select the words in your PDF, it is a scan. Run Searchable before Extract Text, Search or Parser.
- Straighten and rotate before recognition. Sideways or crooked pages give bad text.
- Always set the document language for recognition.
- Recognized text is an estimate. Check names, numbers and legal wording.
- Extraction is not changing. Originals are never modified; you get new files, kept for about two days.
- Do extraction before flattening or locking. Flattened forms and locked files lose or refuse their data.
Continue with how the family works together or try the lab.