PDF to JSON
Map selected fields into JSON.
Extracts the fields you define from one or more PDFs into JSON, with the page each value is on.
How to use it
- Set up AI once in Chat with PDF: connect your own API key, or download a model to this device.
- Choose the PDFs. Under Fields, define the fields here, or choose paste a JSON Schema and paste one. Your fields are remembered in this browser for next time.
- Select Extract data, then review each document’s values beside its pages and correct them.
- Download JSON (one document, or all of them in one file), a CSV with a row per document, or the JSON Schema the data follows.
Options
| Option | Values | Default |
|---|---|---|
| Documents | Choose two or more files | |
| Describe the fields with | One of Fields to collect, A JSON Schema I paste | Fields to collect |
Name each field, choose its type (text, number, whole number, date, yes or no, currency code, or a list of text) and describe it, or paste a JSON Schema. Your fields are remembered in this browser.
JSON Schema support
The schema must describe an object. Its properties can be:
string, with"format": "date"for dates (writtenYYYY-MM-DD),enumfor a fixed list of values, andpattern,minLengthandmaxLengthnumberorinteger, withminimumandmaximumbooleanarrayof any of those, or of objects whose properties are any of those (such as line items)
A type may include null. Objects inside objects, $ref, oneOf and allOf aren’t supported: superPDF says which property to change. Descriptions and titles are passed to the AI to explain each field. Up to 40 fields.
The JSON you get
For each document: file, data (your fields, every value either of its type or null), and citations, which mirrors data with each value’s page, the quote that states it, and its status: found, corrected (found on another page than the AI cited), unsupported (not found in the document, so null), invalid, empty or edited (changed by you). Export waits until every value matches its field.
How values are checked
As in Invoice data extraction: every value must be on a page the AI read, matched as words, as a number however it is written, or as a date in any common form. A value that isn’t there is left null and its suggestion is shown, with a button to use it anyway. Yes-or-no values can’t be matched by their value, so the words the AI quotes for them must be on the page.
Limits
superPDF sets no file size or page limit. Long documents are read in part: with a model on this device, about 7,000 characters (the passages that best match your fields); with your key, about 60,000. Values that must be worked out rather than read, such as a total of several figures or an age from a date of birth, are marked as not found. With your API key, each document’s text goes from your browser to your provider.
Scanned pages
Scanned pages, including copier scans whose text layer is weak or garbled, are read with the quick reader first (the AI tools run OCR PDF for you), so you get a result straight away. On computers that can run it, superPDF then offers high-accuracy reading beside the result, and improves those pages in the background: in Chat and Ask multiple PDFs the pages behind your latest question come first and new questions use the improved text; elsewhere, Run again with the improved text redoes the result. It runs on this device: nothing leaves it, and it works offline once downloaded. Download: 871 MB, once, only when you choose it; delete it any time in On-device AI models. Speed: about 10 to 30 seconds a page on a recent computer (the offer shows the time for your pages); up to 20 pages are read at a time unless you choose more or pick pages. It reads English and other European languages best; check results in other scripts, and like any OCR it can misread faded or handwritten text. Running page headers, footers and page numbers are left out of the text. If anything goes wrong, the quick result stays, with a short note. See High-accuracy reading.
Privacy
Runs in your browser, with the AI you choose. With a model on this device, nothing leaves it. With your own API key (BYOK), the document text the tool needs goes from your browser straight to your provider, never through superPDF. See Privacy: AI features.