Paperwise helps you OCR, extract, organize, and query documents on your own infrastructure. Run it locally or self-host it, and keep full control of your data.
| Name | Date | Status |
|---|---|---|
| Invoice_Dec_invoice_I.pdf | March 12, 2025 | parsed |
| All orders from the records | ||
| Monthly Payments | September 1, 2024 | |
| Paperwise Due | See related results |
Paperwise combines OCR, extraction, organization, and grounded Q&A in one self-hosted workflow.
OCR & extraction accuracy
Go from a 50-page PDF to structured data in seconds. Paperwise can switch between local OCR and LLM-based OCR depending on document quality, helping with scans, dense layouts, and messier files.
| Name | Date Received |
|---|---|
| Sonic Invoice | March 12, 2025 |
| Annual Financials | September 1, 2024 |
| Auto Rate | April 5, 2025 |
Flexible document categories
Auto-tag documents by type, date, entity, and custom taxonomy. Build filterable views across your library and refine the structure over time.
Source-backed answers
Ask natural language questions and get answers grounded in your documents, with source quotes you can trace back to the original files.
"What is the billing cap in the Renfroe contract?"
The aggregate liability is capped at $1,000,000 as stated in Section 4, Paragraph 3 of the executed agreement.
3 configurable model slots
Configure different AI models for OCR, extraction, and Q&A. Use fast models for bulk processing and premium models for high-stakes analysis.
Example model configuration
Real queries. Real documents. Real answers with citations.
Query
"What domains I'm paying for on Cloudflare?"
Returned a structured table of every domain, billing period, and cost across all invoices — each row citing its source document.
Across 8 invoices
Query
"How much did I spend on Sonic internet in each month?"
Extracted a 10-month billing history with amounts and service periods, plus a summary of the charge range. Every row cites its source invoice.
Across 10 monthly invoices
Query
"List every time the target rate changed from Federal Reserve"
Identified three rate changes across policy statements and produced a timeline with old ranges, new ranges, and reserve balance rates.
Across 3 policy statements
Self-host Paperwise and turn raw files into structured, answerable knowledge.
Drag and drop PDFs, scans, images, or text documents into your workspace.
OCR runs automatically. Metadata is extracted, and documents are organized.
Use grounded AI to search language across any set of documents instantly.
Clone the repo, run it locally or on your own infrastructure, and keep your document workflows under your control.
Open source and self-hosted
