The platform

AI for medical imaging,
on managed cloud GPUs.

Nalvera is a streamlined platform for medical imaging research. Upload your scans, pick a peer-reviewed model, and run inference on managed cloud GPUs — segmentation, classification, generation, and more. Reproducible results, no infrastructure overhead.

What you get

Managed GPU pool

Submit multi-file jobs and a GPU worker picks it up in seconds. No queue tooling to install, no Python or CUDA versions to match.

Production-grade models

Peer-reviewed AI models in medical image segmentation, classification, and generation

Per-job control

Configure each run in seconds: choose a task or submodel and assign it to a personal or lab budget.

Your data, your rules

We don’t train on your data by default. Research sharing is explicit opt-in. Delete is one click.

EU-hosted cloud infrastructure

Compute, storage, and database stay inside the EU.

Shared lab budgets · optional, contact us

Top up once, invite colleagues, run jobs against a pooled balance. Per-member spending caps included.

Memberships explained

Guest account

No payment details required. Start with a handful of starter credits and top up further whenever you need — credits never expire.

Top-up only Buy as you go

Why become a member

Memberships add a recurring monthly quota and unlock following perks:

  • Discount on annual billing
  • Save 20-32% on top-up credits vs. the €0.25 standard price, depending on your tier
See Plans & Pricing →
Our credit system explained

Two types of credits

You can spend either type on any job.

Membership Monthly quota

Comes with your plan, and resets each billing cycle.

Top-up Buy as you go

Top up whenever you need extra capacity. Credits don't expire — your balance only drops when you actually submit a job.

A typical job - Example

A whole-body CT segmentation runs around 2–3 credits — that's €0.50–€0.75 at the €0.25 standard price, or as little as €0.34–€0.51 on the highest member tier. Smaller ROIs and single-slice jobs cost a fraction of a credit.

Cost scales with image size Job fails for an unknown reason? This triggers an automatic refund to your balance — no support ticket needed.

The exact price is calculated from your scan's voxel count and shown live in the submit form — before you confirm. No surprise bills.

Our partnership program

Distribute your AI model

For academic groups, labs, and vendors with a peer-reviewed model. Publish, get cited, and get compensated using our platform.

We handle the rest

Provide us with your model (weights). We take care of everything around it:

  • Hosted GPU compute — no servers to procure
  • Flat per-job fee in EUR, set per (model / submodel) link
  • Version-pinned containers for reproducibility
  • Usage analytics, citation tracking, and a public leaderboard
  • User support, billing, and GDPR / DPA paperwork
Get in touch →
How it works
1

Upload your image(s)

Drag-drop a NIfTI .nii.gz file up to 250 MB. We validate the shape and content before queuing.

2

Pick an AI model

Browse the model catalog and pick what fits your image — segmentation, classification, generation, or analysis. The cost preview updates live.

3

Submit & Monitor

A worker picks up your job. Watch progress live on the Job Queue, with expected runtime and status updates.

4

View & Download Refine the result. Segmentation outputs can be fine-tuned directly in your browser with Nalvera's in-house correction toolkit — including an adaptive brush that follows intensity boundaries for fast, accurate edits, plus paint, erase, and slice-by-slice tools. Changes save back to the NIfTI file on disk, preserving the original affine and header.

Open the inference result in the built-in image viewer and/or download the NIfTI / JSON output — ready for ITK-SNAP, 3D Slicer, FIJI, or your own pipeline.

Create a free account Browse models
Return to Platform