Convert Handwritten PDFs to Excel

Handwriting OCR turns handwritten tables, logs and repeated forms into structured spreadsheet data without requiring you to type every row yourself.

There are two routes:

• Tables reads rows and columns from a handwritten or printed table and exports the result as Excel (.xlsx) or JSON.
• Custom Extractors pull the fields you choose from repeated forms and export one row per document as Excel, CSV or JSON.

Test your document with free trial credits. No card is required. Tables and Custom Extractors can be tested with trial credits; continued access requires a Pro or Business plan.

Choose the output that matches the document

This distinction matters. A full-text transcription can preserve the words on a page, but it does not turn a form into one clean spreadsheet row. If every document asks for the same information—name, date, reference number, total, category—a Custom Extractor is usually the better route.

How to convert a handwritten PDF to Excel

1. Choose a representative sample. Include the handwriting, layouts and scan quality that appear in the real batch. For repeated forms, include examples with blank fields, corrections and unusually long answers.
2. Create an account and upload the PDF. Dashboard uploads accept PDFs and common image formats, with files up to 100 MB.
3. Choose Tables or a Custom Extractor. Use Tables for an existing grid. For forms, define the fields you need; the extractor can suggest a schema from a sample for you to review.
4. Process and inspect the result. Each processed page uses one credit. Check names, dates, totals and other critical fields against the source before importing the file elsewhere.
5. Download the spreadsheet. Tables export to XLSX. Custom Extractors can export to XLSX, CSV or JSON.

Start with your own sample

Handwritten tables: keep the rows and columns

Use Tables when the source already has a tabular structure: inventory sheets, field observations, attendance lists, handwritten ledgers or laboratory tables.

The service processes multi-page PDFs page by page and returns structured table data for download. Results depend on the legibility of the writing and the consistency of the layout, so review a representative sample before committing a large collection.

For page-size documents, scan at about 300 DPI in greyscale or colour. Higher resolution is rarely useful once the page is clear and human-readable. If you are starting from a scanner, the scan-to-OCR guide explains the settings and email-submission workflow.

Repeated forms: choose the fields you need

Use a Custom Extractor when every document follows a similar form but you only need particular fields. Define columns such as:

• respondent name
• date
• location
• reference number
• amount
• category
• notes

The result is one spreadsheet row per processed document instead of a block of transcription text. This is useful for surveys, registrations, applications and operational paperwork that must feed another system.

A Virginia insurance agent used a Custom Extractor for 50–80 handwritten lead forms each week, replacing 2–3 hours of manual entry before the data reached a spreadsheet and CRM. Read the customer story.

A real handwritten-form workflow

Project Belize processes roughly 700 handwritten clinic intake forms collected during medical missions. The forms arrive folded, marked and sometimes difficult to read. Their team now extracts the data into Excel for its health-ministry report instead of asking five or six volunteers to key in names for several hours each evening.

The customer reported accuracy above 95% on its documents and at least an 80% reduction in labour. Those figures describe that project, not a guarantee for every form. Read the Project Belize story.

Test accuracy before processing the whole batch

Handwriting, scans and layouts vary, so there is no useful universal accuracy percentage. Test pages from the difficult end of your collection and check the fields that matter most.

For an operational workflow:

• verify totals and identifiers against the source
• include blank, crossed-out and faint examples in the sample
• agree how missing or uncertain values will be handled
• spot-check batches after the initial test
• keep the original documents available during review

If a job is large enough to need file preparation, managed delivery or a defined acceptance process, use the managed transcription service rather than forcing it through a self-service workflow.

Plans, files and privacy

New accounts receive five trial credits and do not need a card. Trial credits can be used to test Tables and Custom Extractors on the documents you actually need to process. Continued access to these features requires Pro or Business; one page uses one credit.

Dashboard uploads accept PDF, JPG, PNG, GIF, HEIC and TIFF files up to 100 MB. For large sources, split files into manageable PDFs of roughly 100 pages.

Customer documents are not used to train models. Files and extracted results are encrypted in transit and at rest. The default automatic deletion window is seven days and can be configured from 15 minutes to 14 days; you can also delete documents manually.

Start with one representative document

Choose a sample that reflects the handwriting and layout in the full batch. Use Tables for an existing grid or a Custom Extractor for repeated named fields, then inspect the spreadsheet before scaling up.

Test a handwritten PDF with free trial credits

Tables and Custom Extractors can be tested with trial credits. Continued use requires Pro or Business.