napari Chat Assistant

napari Chat Assistant brings local AI into the napari image viewer for researchers, imaging facility staff, and educators working with microscopy and multidimensional datasets. It uses the layers and regions of interest already open in your session to interpret requests and carry out analysis. Image data stays on your workstation.

The plugin runs open-weight models through Ollama, including nemotron-cascade-2, qwen3.5, and gpt-oss. Normal use doesn't need an internet connection or API key. It's open source under the MIT license; Ollama and model weights are separate requirements, and larger models can need substantial RAM or VRAM.

Built-in actions cover image enhancement, thresholding, mask cleanup, object measurement, projections, and cropping. You can also extract values from regions of interest, compare groups, inspect intensity profiles, and add annotation overlays without changing the underlying image. These actions can run directly, without relying on a model to interpret a prompt.

For custom analysis, the assistant generates Python tied to the current napari viewer. You can review and execute that code inside the plugin, or ask it to explain and repair your own Python. Generated code can still contain incorrect APIs or imports, so it needs review.

A local library holds reusable prompts, code, and templates, while saved shortcuts turn repeated tasks into buttons. Workspace saving preserves session state and generated image or label data in OME-Zarr assets. Teaching tools include synthetic datasets, and optional voice input transcribes speech locally with faster-whisper.