An open-source, modular digitization toolkit designed for low-cost, high-quality scanning using Raspberry Pi Compute Module 4 and 64MP ArduCam cameras. Built for use in low-resource environments and community archives.
- Empower small institutions and communities to digitize archival materials.
- Provide an easy-to-use web interface for image capture, metadata entry, and export.
- Ensure long-term reproducibility through open standards (BagIt, CSV, TIFF).
- Prioritize modularity and hardware independence for future-proofing.
- Hardware: Raspberry Pi CM4 + CM4IO, Raspberry Pi 5 (dual camera embedded), 2x 64MP ArduCam autofocus cameras, 2 x Cannon EOS Rebel T7 (GPhoto compatible cameras)
- Backend: Python, FastAPI
- Frontend: Svelte
- Database: PostgreSQL
Project currently in development. Kick-off: September 2025 Alpha prototype planned for deployment at SBMAL in June 2026.
See CHANGELOG.md for version history and release notes.
# One-command startup (all services in Docker)
./scripts/start-dev.shOr manually:
docker compose -f docker-compose.yml -f docker-compose.dev.yml --profile with-backend up# Build the frontend image first (requires internet — do this at home/office):
docker compose build
# Then take it offline and run:
./scripts/start.shOr manually:
# Start database and frontend
docker compose up -d
# Start backend with pixi
cd backend && pixi run devAccess (production): http://localhost:3000 (frontend) | http://localhost:8000/docs (API) Access (dev): http://localhost:5173 (frontend, Vite) | http://localhost:8000/docs (API)
Note: The native backend is required for camera access due to Raspberry Pi-specific libraries (libcamera, picamera2).
The backend uses pixi for dependency management:
# First time setup
cd backend
pixi install
pixi run setup-camera-link # Raspberry Pi only
# Start backend
pixi run devThe toolkit is designed to ship as a pre-flashed SD card. The end user only needs to insert the card, power on the Pi, and the application starts automatically — no internet connection required.
# 1. Clone the repository onto a fresh Raspberry Pi OS installation
git clone --recurse-submodules https://github.com/UCSB-AMPLab/digitization-toolkit.git ~/dtk
cd ~/dtk
# 2. Run the one-time provisioning script (internet required here)
# Builds Docker images, installs pixi env, applies DB migrations
sudo ./scripts/setup.sh
# 3. Install and enable the systemd service (auto-start on boot)
sudo ./scripts/install-service.sh
sudo systemctl start dtk
# 4. Verify the app is running
curl http://localhost:8000/health # → {"status":"ok"}
curl -I http://localhost:3000 # → HTTP 200# On the Pi — shut down cleanly
sudo systemctl stop dtk
sudo poweroff
# On a host machine with the SD card inserted (replace sdX with your device)
sudo dd if=/dev/sdX of=dtk-$(date +%Y%m%d).img bs=4M status=progress
# Compress for storage/transfer
xz -T0 dtk-$(date +%Y%m%d).imgUse Raspberry Pi Imager (or dd) to write the .img to a new SD card.
The application starts automatically on boot — no configuration needed.
| Component | Stored at | Requires internet after setup? |
|---|---|---|
| Frontend (SvelteKit) | Docker image in /var/lib/docker/ |
No |
| PostgreSQL | Docker image in /var/lib/docker/ |
No |
| Backend Python env | backend/.pixi/ |
No |
| camera libraries | /usr/lib/python3/dist-packages/ |
No (system OS) |
| Application data | /var/lib/dtk/ |
No |
This repository uses Git submodules for the frontend and backend code.
When cloning, make sure to fetch submodules as well:
# Clone with submodules
git clone --recurse-submodules [email protected]:UCSB-AMPLab/digitization-toolkit.gitIf you already cloned without --recurse-submodules, you can initialize and update submodules manually:
git submodule update --init --recursiveTo pull the latest changes for submodules after updates:
git submodule update --remote --mergeSee the wiki for detailed developer guides and API references.
MIT