Source-cited document extraction

Structured data with a paper trail

0files kept after extraction
5documents free, no credit card required

Every figure linked to the page it came from

The principle

Try a sample document

Interactive demo

See where each value came from, ask the Inspector, and see what a reviewer should check.

app.validextract.com/workspace

Choose a template

Upload your document
10-K / Financials · Eargo, Inc. 10-K, fiscal 20229 fields · 3 flagsCited 9 of 9

Eargo, Inc. annual report on Form 10-K for the year ended December 31, 2022; Eargo is also the tenant in our commercial lease sample. It’s a public document, filed on SEC EDGAR. It is not a client document.

Fiscal yearYear ended December 31, 2022
Net revenue$37.2 million, up $5.1 million from 2021
“For the year ended December 31, 2022, we generated net revenue of $37.2 million, an increase of $5.1 million from the year ended December 31, 2021.”Source: p. 57, Item 7
Systems shipped24,247 in 2022, down from 45,136 in 2021
Gross margin38.3%, up from 13.0% in 2021
Net loss$157.5 million (2021: $157.8 million)
Cash$101.2 million in cash and cash equivalents
Pricing concession$11.6 million reduction to net revenue for unsubmitted and unpaid insurance claims
Going concernSubstantial doubt about the ability to continue as a going concern
EmployeesAbout 243 full-time, 236 in the United States

What the document says, not what the model guessed. Every value opens the page it was read from.

Reviewer flags

  • Net revenue rose $5.1 million while systems shipped fell from 45,136 to 24,247. Read the revenue reconciliation on p. 57–58 before comparing the two years.
  • 2022 revenue already includes an $11.6 million reduction from the insurance pricing concession (p. 58), so it is not a clean run-rate figure.
  • The filing raises substantial doubt about the company’s ability to continue as a going concern (p. 14). Carry that into any model built on these figures.

What it is

A verbatim-grounded document extraction platform for institutional due diligence, financial analysis and legal workflows. Dense filings and agreements become clean, structured fields.

See the features

Who it is for

Boutique M&A advisory, private equity deal teams, corporate legal counsel and equity research analysts. Anyone who has to defend a number back to the page it came from.

Source-anchored values

Each value is tied to a page and paragraph. If the text is not there, the field is left blank.

Zero-retention security

Uploads never train a model and are cleared from memory when extraction finishes.

1-click templates

Prebuilt extractors for annual reports, diligence checklists and contract review. No prompts.

Audit-ready exports

A branded PDF with your logo and clickable citation footers, plus CSV and JSON.

How it works

new-extractionHTTPS
PDF
annual-report-2025.pdf
112 pp
Template10-K financials
Fields to extract18
Prompts to writeNone
Model to trainNone
Template matchedReady
01 · Upload

Drop a filing, it finds the template

Upload a PDF, docx or xlsx through the encrypted dropzone. The document type is detected and matched to a template, so there is nothing to set up.

Encrypted dropzone · up to 20 MB
extract · 10-K1440 · reading
Total revenue $48.2Mp. 7
Gross margin 52.4%p. 7
Net income $6.1Mp. 9
Long-term debt $22.5Mp. 31
Renewal optionBlank
18 fields18 cited
02 · Extract

Every value lands with its page

Each figure and clause is tied to the page and the exact line it was quoted from. If the text is not in the document, the field stays blank instead of guessed.

Page + quote on every value
review · exportRound 1
Revenue growth rounds 11.1% to 11%Flag✓ Checked
Share count is in a later noteFlag✓ Checked
Values cited18 of 18
Branded PDFCSVJSON
2 flagsAll checked
03 · Verify

Check the flags, then export

Reviewer flags point at anything missing, unclear or inconsistent. Click a value to see its source, then export to CSV, JSON or a branded PDF with citation footers.

PDF · CSV · JSON

Industries

Drag to move

Certified and compliant

ValidExtract is built on ISO 27001 and SOC 2 Type II audited infrastructure and is GDPR-ready. Files are purged the instant extraction completes, success or failure, and never train a model.

  • Files stored after extraction0
  • Documents free to start5
  • Export formats, each with citations3
ISO
27001
ISO 27001Information Security Management
SOC 2
SOC 2 Type IIService Organization Controls
GDPR
GDPRData Protection & Privacy

On every plan, for every file.

  • Zero-RetentionFiles are deleted when extraction completes.
  • Encrypted in TransitUploads and results travel over HTTPS.
  • HIPAA-ReadyPHI is extracted and purged per request, not retained.
  • SOC 2 Type IIHosted on SOC 2 Type II audited infrastructure.
  • Never Trains a ModelDocuments are not used for training or tuning.
  • GDPR-ReadyDeletion and data-subject rights are supported.
  • Encrypted DropzoneUploads enter through an encrypted channel.
  • Cited, Defensible ExportsEvery field carries its source citation.

FAQ

Answers to the key questions

01Is there a free plan?

Yes. The Free plan is $0, needs no card, and covers 5 documents (up to 200 pages in total) with every template, source citations and zero-retention processing. You can also run the sample document on this site without signing up, or email us up to 5 of your own documents and get them back cited within 48 hours, free.

02Which documents can I upload?

Long financial, legal and real-estate documents: 10-K and 10-Q filings, annual reports, CIMs, purchase agreements, contracts and commercial leases. There are also templates for invoices and vendor statements, insurance policies and loss runs, RFPs and tenders, construction subcontracts and clinical study reports. Upload a PDF, docx or xlsx file of up to 20 MB; text PDFs and scans both work.

03How do source citations work?

Every extracted value carries the page and the passage it was read from. Click a value and the document preview jumps to that line, highlighted, so a check takes seconds rather than a search through the file.

04What happens when a value is not in the document?

The field stays blank and is flagged. ValidExtract never fills a gap with a guess. Reviewer flags also point out anything unclear or inconsistent, such as a figure that is defined differently in two places, so you know exactly what to look at before you rely on the output.

05What happens to my files?

They are processed in memory and deleted the moment extraction finishes, on every plan including Free. Nothing is stored afterwards and your documents are never used to train a model.

06Do I need to write prompts or train a model?

No. Upload the file, the document type is matched to a ready-made template, and the fields come back with their citations. There is nothing to install and nothing to configure; three steps from upload to a checked export.

07How is a document counted, and what do the plans cost?

A document is one file of up to 200 pages; a longer file counts as one document per 200 pages. Free: 5 documents, up to 200 pages in total. Pro: $299 a month for 50 documents. Team: $499 a month for 125 documents and 3 seats. Corporate: $3 per document with a one-time $4,997 setup fee and no monthly minimum. Prices are in US dollars.

08Can I export the results?

Yes. Export to CSV, JSON, or a branded PDF whose citation footers link back to the source pages, ready to hand to a partner, a lender or an auditor.

09How is ValidExtract different from other extraction tools?

Every value is tied to the exact page and passage it came from, and a field is left blank rather than guessed when the text is not there. It is built for people reading long filings, agreements and leases, needs no prompts or training, and does not keep your files. See how it compares with other tools.

10Which security and compliance standards do you meet?

ValidExtract is built on ISO 27001 and SOC 2 Type II audited infrastructure and is GDPR-ready, with encryption in transit over HTTPS.

No sign-up

Try a sample document

See every value and the page it came from. Free for your first 5 documents.

Run the sample