Paper-heavy workflows to usable data

Turn messy documents into clean business outputs.

DocUnlocked helps logistics, construction, property, bookkeeping and technical teams convert scans, forms, tables and PDFs into CSV, Excel, Word, Markdown, JSON and AI summaries.

Test 5 pages See proof library

Review before useUncertain fields stay visible.

Source traceabilityExamples cite public documents.

No API setup firstInterface now; API later.

Real handwriting transformation

1. ImportPDF, image, web or office file

2. DetectLayout, tables, formulas, figures

3. RebuildReadable document + structured data

4. ExportCSV, Excel, Word, JSON, ZIP

NIST handwriting formReal handprinted fields and paragraph

Real cursive lab notebook page from Library of Congress

Lab notebookCursive, faded ink, experiment notes

Real handwritten 1930 census table from the National Archives

Handwritten tablePublic National Archives schedule

Public GSA Government Bill of Lading form preview

Freight documentPublic GSA bill of lading form

NIST handwriting review packagePublic source

Review package preview for a real NIST handwriting form

Real handwriting shown
The source is a public benchmark image, not a staged handwriting sample.
Review-first output
Low-confidence fields remain visible before export.
Structure preserved
Boxes, rows, sections and source crops stay attached.
Business exports
CSV, Excel, Word, Markdown, JSON and summaries are the deliverable.

Clean spreadsheet rowsReady to review

field	value	confidence
document_type	Technical report / field packet / form	High
tables	Detected table regions exported as rows	High
figures	Image and chart blocks preserved	Review
formulas	Formula-like regions extracted or preserved	Review
exceptions	Handwriting, low contrast, damaged scans	Flagged

Workbook tabsOps-ready

Summary - document type, source, page count, extraction status.
Fields - normalized key/value rows with review flags.
Tables - shipment lines, inspection deficiencies, invoice lines or lab measurements.
Review - low-confidence fields and handwritten notes.

Editable recovery documentReview required

Recovered headings and sections.
Tables preserved as editable structures when possible.
Images, formulas and complex regions kept as traceable blocks.
Useful for teams that need a human-readable final file, not only CSV.

Structured technical payloadAPI-ready

{
  "workflow": "complex_document_recovery",
  "outputs": ["markdown", "docx", "xlsx", "html", "json", "summary", "zip"],
  "review_flags": ["formula_region", "figure_caption", "handwriting"],
  "source_trace": "public document + page coordinates"
}

Operator summaryFast triage

Document contains multi-column text, tables, figures and formula-like regions.
Clean sections and structured tables can feed review, reporting or downstream AI workflows.
Best next action: validate low-confidence figures/formulas before official reuse.

MarkdownAI-ready structure

XLSXTable workbook

DOCXEditable recovery

ZIPFull package

Industry proof library

Lead with the industries still buried in paper.

The buying question is not "does OCR exist?" It is "will this work on my documents, in my workflow, without an engineering team?"

Logistics

Bill of lading to CSV

Extract shipper, consignee, BOL number, commodity rows, weights, charges and delivery exceptions from freight paperwork.

View use case

Construction

Field reports to Excel

Turn daily reports, stormwater logs, payroll packets and supplier forms into workbook tabs for project controls.

View use case

Property

Inspection checklist to spreadsheet

Extract unit, deficiency, severity, area, owner notes and action items from inspection forms and PDF checklists.

View use case

Labs / engineering

Technical PDFs to Markdown + JSON

Recover tables, formulas, figures, sections and references for bureau d'etude, lab and research teams.

View use case

Public-source proof

Real handwriting first, with reviewable outputs.

The site now shows public handwritten sources with visible transformation steps. The promise is not magic handwriting OCR; it is a clean review package the operator can validate and export.

Real handprinted form

NIST SD19 handwriting sample form

A public NIST handprinted form shows real field boxes, digit strings and a handwritten paragraph. This is the right proof class for teams with forms completed by hand.

HandwritingFieldsReview flagsCSV

Source: NIST SD19 handwriting sample image. Output shown as a DocUnlocked review/export package preview.

BeforePublic image

AfterReview package

DocUnlocked review output preview for NIST handwriting form

Real handwritten table

1930 census schedule to rows

Dense handwritten tables are where simple document demos collapse. The value is row alignment, columns, review flags and spreadsheet-ready output.

TablesRowsExcelReview queue

Source: U.S. National Archives 1930 census schedule PDF. Output shown as a review/export package preview.

BeforePublic PDF

Real handwritten 1930 census schedule table

AfterRows + review

DocUnlocked rows and review preview for handwritten census table

Real cursive lab notes

Lab notebook to Markdown + review queue

Labs, bureaus d'etude and archives need more than CSV. Cursive notes should become Markdown, searchable text, source crops and uncertainty flags.

CursiveMarkdownSource cropAI summary

Source: Library of Congress Alexander Graham Bell notebook page. Output shown as a review/export package preview.

BeforePublic image

AfterMarkdown + review

DocUnlocked Markdown and review preview for cursive lab notebook

Public government PDF

IRS W-4 / vendor-admin style forms

Public tax forms prove structured field and table recovery. For live customers, this maps to W-9, onboarding packets and vendor files.

FieldsTablesWordJSON

Source: irs.gov W-4 PDF. Public federal document processed through MinerU for the deployed proof asset.

BeforePDF

AfterMinerU preview

After screenshot of extracted IRS W-4 output

Technical documents

Tables, formulas and multi-column reports

For labs, engineering offices and technical teams, the final deliverable is often Markdown, DOCX, HTML, JSON and a structured summary, not just a CSV.

TablesFiguresFormulasMarkdown

Synthetic technical stress sample processed through the real MinerU flow. No private or customer document used.

BeforePDF

Before screenshot of technical PDF stress sample

AfterMinerU preview

After screenshot of technical PDF extraction output

SEO attack surface

Own the long-tail search intent by document type.

Every vertical page targets a concrete pain: paper document in, usable output out.

Freight documents

bill of lading to CSV, freight invoice to Excel, proof of delivery OCR.

Bill of lading page

Construction paperwork

daily field report to Excel, stormwater inspection PDF extraction, contractor forms to spreadsheet.

Construction page

Property operations

inspection checklist to spreadsheet, NSPIRE checklist to CSV, maintenance forms to action list.

Property page

Bookkeeping and vendor admin

W-9 to CSV, vendor forms to spreadsheet, scanned invoice lines to Excel.

Vendor forms page

Labs and bureau d'etude

technical report to Markdown, tables and formulas to JSON, PDF figures to AI-ready package.

Lab reports page

Handwritten paperwork

handwritten forms to Excel, handwritten table to spreadsheet, handwriting OCR with review flags.

Handwriting page

General scanned forms

scanned forms to Excel, PDF table to CSV, paper forms to structured data.

Scanned forms page

North star product path

Interface first. Chrome extension next. API when volume proves it.

Web proof lab

Users upload 5 pages and see the output before trusting the system.

Watched inbox

Forward PDFs from AP, dispatch or property inboxes into a review queue.

Chrome extension

Capture PDFs from portals and web apps without making users download/re-upload manually.

Workflow exports

Push CSV/XLSX/JSON/Markdown packages into Sheets, Airtable, accounting tools or ERPs.

API

Offer API access only after repeatable vertical workflows and willingness-to-pay are proven.

Test your own document

See what comes back before betting the workflow on it.

Upload a file you are authorized to process. The launch flow validates the file and prepares a secure checkout or support handoff depending on backend availability.

Launch pricing

Simple enough to test, capped enough to protect margins.

Single

Single Document Unlock

One document, up to 200 pages or 200 MB.

Pack

5 Document Pack

$29

Credit pack for multiple archives, forms or contracts.

Power

20 Document Pack

$79

Small business pack for repeated document recovery.

Privacy and handling

MinerU and payment secrets stay server-side in the backend design.
No private customer, Benjamin or F2M documents are used in public proof assets.
Cloud parsing is used; regulated or highly sensitive documents require the customer's own approval path.
Results must be reviewed before legal, financial or official use.

Accuracy note.
DocUnlocked is a document recovery and workflow extraction service, not a certified transcription or legal records service. Scan quality, handwriting, layout, language and file condition affect results.