fn0000/fn-data-server repository overview

⁠Find Nuclei Data Server

On-site S3-compatible HTTP server for streaming OME-ZARR⁠ microscopy data to the Find Nuclei Viewer⁠.

Deploy on your lab network, NAS, or HPC cluster. The browser fetches ZARR chunks directly from this server — your data never leaves your infrastructure.

🔬 Viewer: find-nuclei.github.io⁠

⁠Quick Start

mkdir -p ~/.findnuclei

docker run -d \
  -p 3333:3333 \
  -p 3334:3334 \
  -v /path/to/your/zarr/data:/data/storage \
  -v ~/.findnuclei:/data/config \
  -e FN_ADMIN_PASSWORD=changeme \
  fn0000/fn-data-server

• Data server: http://localhost:3333⁠ — S3-compatible API for ZARR chunks
• Admin console: http://localhost:3334/admin⁠ — token management, bucket browser

⁠View Your Data

Open the viewer with a direct URL to your data:

https://find-nuclei.github.io?url=http://localhost:3333/my-bucket/sample.zarr

If auth is enabled, add your token:

https://find-nuclei.github.io?url=http://localhost:3333/my-bucket/sample.zarr&token=YOUR_JWT

⁠Architecture

┌──────────────────┐     ┌──────────────────┐     ┌──────────────────┐
│  Find Nuclei     │     │  User's Browser  │     │  Data Server     │
│  Platform        │◄───►│  (Viewer)        │◄───►│  (Your Network)  │
│  metadata only   │     │                  │     │  ZARR chunks     │
└──────────────────┘     └──────────────────┘     └──────────────────┘

The browser is the bridge. The data server only needs to be reachable from the user's browser (corporate network, VPN). The cloud platform never connects to it directly.

⁠Split Ports

⁠Operation Modes

⁠Secure Mode (default)

All data endpoints require a JWT Bearer token. Create tokens via the admin console and scope them to specific buckets.

docker run -d \
  -e FN_AUTH_ENABLED=true \
  -e FN_ADMIN_PASSWORD=mypass \
  ...
  fn0000/fn-data-server

⁠Open Mode

No authentication on data endpoints. Use only on trusted networks.

docker run -d \
  -e FN_AUTH_ENABLED=false \
  ...
  fn0000/fn-data-server

⁠Token Scoping

Each JWT token is scoped to a bucket prefix:

• my-project — access only objects under /my-project/
• * — access all buckets (wildcard)

Tokens also support:

• TTL — automatic expiration
• Host restriction — limit to specific domains (supports wildcards: *.lab.internal)
• IP restriction — limit to specific client IPs (supports wildcards: 192.168.1.*)

⁠Environment Variables

⁠Data Setup

Place your OME-ZARR datasets in the storage directory. Each top-level directory becomes a bucket:

/your/zarr/data/
├── project-alpha/
│   ├── sample1.zarr/
│   └── sample2.zarr/
└── project-beta/
    └── experiment.zarr/

The admin console's bucket browser shows all top-level directories. Click a bucket to pre-fill the prefix when creating a token.

⁠Scaling

• 1-5 users: Default config (4 workers) handles this comfortably
• 5-20 users: Increase workers: -e FN_DATA_WORKERS=8
• 20-50 users: Multiple containers behind a load balancer
• 50+ users: Contact us about Enterprise deployment

⁠API Endpoints

⁠Data Server (port 3333)

⁠Admin Console (port 3334)

⁠Platform Integration

1. Start the data server on your network
2. In Find Nuclei Platform → Settings → Data Sources, add your server URL and token
3. Click "Test Connection" to verify
4. Add images using paths like: http://your-server:3333/project-alpha/sample1.zarr

⁠Links

• 🔬 Free Viewer: find-nuclei.github.io⁠
• 🌐 Platform: find-nuclei.github.io⁠
• 📖 OME-ZARR Spec: ngff.openmicroscopy.org⁠

⁠License

MIT