Back to all projects
Computer Vision2025Featured

SignKit

Signature extraction & PDF signing — commercial desktop product

PySide6FastAPIPostgreSQLOpenCVStripePythonPyInstaller

Problem

Legal and financial teams process thousands of signed documents daily. They need signature extraction, PDF signing, and organized output management in one complete workflow.

Approach

PySide6 cross-platform desktop app with FastAPI backend. PostgreSQL for document metadata and user management. OpenCV-based signature extraction with morphological operations. PDF signing with embedded certificate support.

Technical Implementation

CV Pipeline

OpenCV morphological operations (dilation, erosion, contour detection) to isolate signature regions. Handles handwritten signatures on scanned documents, photos, and mixed backgrounds.

Backend

FastAPI with async document processing. PostgreSQL for user accounts, document metadata, and audit trails. File storage with organized output folders.

Digital Signing

PDF signing with embedded certificate chains. Signature appearance customization. Batch signing for multi-document workflows.

Monetization

Stripe integration for subscription billing. License key validation. Trial mode with full feature access for 14 days.

Distribution

Cross-platform builds for Windows and macOS. Code-signed binaries. Auto-updater integration.

Outcomes

  • Shipped as paid product with Stripe subscription billing
  • Signature extraction accuracy >90% on clean documents, >75% on degraded scans
  • Batch processing workflows save legal teams hours per week
  • Cross-platform deployment (Windows/macOS) with code signing

Ownership & scope

Owned product and build from idea to paid release: desktop UX, extraction/signing pipeline, billing integration, packaging, and shipping workflow.

Constraints

  • Needed both signature extraction and signing in one operational flow, not separate tools
  • Cross-platform distribution and update reliability for non-technical end users
  • Accuracy requirements on imperfect scanned documents

Trade-offs

  • Chose desktop-first workflow (PySide6) for local file-heavy operations instead of web-only upload flow
  • Focused initial release on core extraction + signing reliability before broader template and collaboration features

What changed

  • Replaced fragmented manual signing process with a single structured workflow
  • Moved from prototype utility to paid product with licensing and subscription flow

Workflow artifacts

  • Batch signing queue with status tracking
  • Signature region extraction pipeline and confidence checks
  • Stripe-backed subscription and license handling

Result

Commercial signature extraction product built from idea to paid product, including extraction, signing, and workflow tooling.