snappdf

View live

Open-source PDF toolkit — merge, split, sign, scan and convert to Word or Excel — where every file operation runs in the browser, so nothing is ever uploaded.

snappdf landing page — “The ultimate PDF workspace. Zero paywalls.”
snappdf architecture — every tool runs in the browser; a keyless GitHub Actions pipeline deploys the service to Cloud Run
snappdf's browser network panel while signing a document, showing zero requests and 0 B transferred — proof the file never left the browser
snappdf on mobile — the camera scanner auto-cropping a page, and the PWA merging PDFs

snappdf is a free, open-source PDF toolkit — merge, split, sign, scan and convert to Word or Excel — built around one constraint: no file ever leaves the browser. Every operation runs client-side with pdf-lib, pdfjs-dist and docx, including a camera scanner that auto-crops and straightens a photographed page, with HEIC support for iPhone photos. It installs as a PWA with a service worker, so it keeps working after the first load with no connection. I built a keyless CI/CD pipeline that type-checks, tests, audits and builds a Docker image on every pull request, then deploys straight to Cloud Run through GitHub Actions and Workload Identity Federation, with no service-account keys stored anywhere. The repo runs like any other open-source project, with branch protection, Dependabot and a CONTRIBUTING guide for outside contributors. The one thing that does touch a server is the in-app bug-report form, handled by a small Node sidecar scoped to its own secrets with no access to user files.

Technologies

  • React
  • TypeScript
  • Vite
  • Tailwind CSS
  • Vitest
  • Docker
  • GitHub Actions
  • Cloud Run
  • Node.js
Next