A complete medical image viewer, in your browser.

No installer, no server, no account. Drag in a study and it is decoded, rendered and segmented on your own machine — and your annotations, measurements, reports and datasets stay in the browser, or in a folder you choose, rather than in an ImagiVue repository.

Beta access is invitation-only for now. Already invited? Use button above.

Not cleared for clinical diagnostic use.
For research and educational purposes only.

A few of the things you will be testing

A sample of the viewer, not the full tour — the rest is waiting behind the login.

Open Almost Anything

DICOM, NIfTI and multi-frame studies, plus JPEG, PNG and TIFF
Drag in files, or scan a whole folder recursively
Or connect straight to a PACS over DICOMweb

GPU Rendering Built for Reading

Window/level, zoom and pan run on the GPU, not the CPU
Layouts up to 6x4 on WebGPU, with MPR and cine playback
Automatic WebGL2 fallback and device-loss recovery

In-Browser AI Segmentation

Click one point and MedSAM proposes the region
Boundary refinement snaps the contour onto the real image edge
No upload, no GPU server, no queue

Browser-Local Imaging by Design

Don’t take our word for it — open DevTools and watch the Network tab while you work.

✓Intended viewer workflows keep images browser-local — verify ordinary traffic in your DevTools Network tab
✓AI segmentation runs in your browser — no upload, no GPU server, no queue
✓DICOMweb connections you configure go directly from your browser to your PACS, with no proxy in between. The built-in demonstration connection, “Orthanc Demo (UCLouvain)”, is the exception: its requests go to the website that serves the viewer, which relays them to a public demonstration server. It is read-only, and the viewer labels it “Public demonstration server — do not enter or send patient data”
✓DICOM files are parsed and rendered entirely in the browser
✓Annotations, measurements and reports are stored locally — as plain files on disk once you connect a data directory — with no routine ImagiVue cloud upload
✓The cloud serves application code and AI model weights, and relays DICOMweb requests only for the built-in demonstration connection — no analytics or tracking pixels touch patient data

Cloud (No PHI)
CDN → App Code + AI Model Weights
Auth Service → Account Management Only

↓ Internet ↓

Your Browser
DICOM Parser → WebGPU Renderer
ONNX Runtime Web → AI Models
Annotations → IndexedDB

↓ Direct Connection (bypasses our cloud) ↓

Hospital PACS
DICOMweb ←→ Browser (direct)