Papereg

Paper workflows hide useful data.

Organizations still depend on PDFs, phone photos, scanned packets, signatures, and handwritten records. Re-entering that information wastes time, creates mistakes, and makes reporting difficult. Simple online form builders do not preserve the structure of real documents or help with existing backlogs.

One path from document to usable record.

Papereg combines form design, public and authenticated submission, document ingestion, AI extraction, human review, workflow state, analytics, and export.

• Convert an uploaded PDF into a proposed digital form with detected fields, labels, sections, and layout.
• Create forms from plain-language descriptions and review the proposed structure before creation.
• Fill drafts from text, photos, scans, handwriting, or multi-page documents.
• Process batches of documents while preserving successful results and retrying only failures.
• Build complex forms with tabs, sections, related fields, signatures, files, media, and validation.
• Use industry templates, dashboards, reports, and PDF or spreadsheet exports.

Models propose. Product rules decide.

AI output enters the product as a proposal, not unquestioned truth. Users can review generated fields and extracted values before saving or submitting. Batch processing exposes created and failed records separately so uncertainty is visible and recoverable.

The workspace assistant is role-aware. It can help users query authorized data, create forms, add fields, and fill drafts, but it cannot delete data, approve records, change permissions, or configure integrations.

The unglamorous parts are part of the product.

I owned the product workflows, data structures, AI-provider boundary, ingestion and retry behavior, review states, permissions, dashboards, exports, deployment, and public product presentation. The goal was not a document demo; it was a system that remains understandable when files fail, users disagree with extraction, and different roles need different authority.