Start here¶
If VIPP is not installed yet, begin with Install VIPP. Checksum-verified unsigned installers are the recommended route on Windows and macOS; they create the normal launch shortcuts or application for you.
Already installed? See check for updates.
If someone has asked you to evaluate the alpha on a new Windows computer, use the short field checklist and leave anything you did not try marked not run.
You do not need to write code to use VIPP. Start by opening a complete workflow on synthetic data, learn how to inspect each stage, and only then build or adapt a graph.
A 20-minute route through the manual¶
| Time | Activity | Outcome |
|---|---|---|
| 5 min | Install and launch | VIPP opens inside napari. |
| 5 min | Tour a finished workflow | You can select nodes, inspect outputs, and identify a manual node. |
| 7 min | Build a small workflow | You can add, connect, tune, and save nodes. |
| 3 min | Switch to your data | You understand source choices and the checks required after transfer. |
Learn on the bundled data first
Bundled samples are deterministic, small, and free of privacy or licensing concerns. If a tutorial behaves differently from the manual on a bundled sample, check the documentation version before debugging your microscope file.
What you should understand before batch processing¶
You are ready to run a folder only when you can:
- tell an image, mask, label image, and table apart;
- confirm the axes and physical scale of an input;
- inspect the mask and labels—not only the final table;
- explain the parameters that materially change the result;
- identify representative images that were not used for tuning;
- distinguish the requested compute mode from the CPU/GPU badge showing what actually ran;
- save a workflow and reopen it in the same VIPP release.
If any of those are unfamiliar, the concepts and scientific-practice sections provide the necessary background without assuming Python knowledge.
Optional GPU use is not required for this route. New sessions default to Auto, which can use a reviewed GPU default only when the installed environment and exact workload pass every gate; otherwise it gives an explained CPU decision. Windows users should follow the dedicated CUDA installation guide. Continue to choose and verify CPU or GPU compute only after the workflow is scientifically sensible on representative data. Prefer GPU uses every reviewed eligible accelerator region globally; Custom adds per-node choices and benchmarking.