Getting Started

Installation

uv pip install anybioimage

Optional extras

Extra	Adds	Notes
`bioio`	`bioio`, `bioio-tifffile`	TIFF and OME-Zarr file reading
`contours`	`scipy`	Contour-only mask rendering
`sam`	`ultralytics`	SAM segmentation (PyTorch required)
`all`	bioio + contours	Recommended for most users
`complete`	all + sam	Full install

uv pip install "anybioimage[all]"       # recommended
uv pip install "anybioimage[complete]"  # includes SAM (Python 3.10–3.12)

SAM and Python 3.13+

SAM support requires PyTorch via ultralytics. Use Python 3.10–3.12 for SAM features.

Usage in Jupyter

Just evaluate the viewer in a cell — it renders inline like any other widget:

from anybioimage import BioImageViewer
from bioio import BioImage

viewer = BioImageViewer()
viewer.set_image(BioImage("image.tif"))
viewer

Usage in marimo

Wrap with mo.ui.anywidget() to make the widget reactive:

import marimo as mo
from anybioimage import BioImageViewer
from bioio import BioImage

viewer = BioImageViewer()
viewer.set_image(BioImage("image.tif"))
mo.ui.anywidget(viewer)

Reactive annotations

In marimo, annotation data is reactive. Any cell referencing widget.value will re-run automatically when you draw or delete annotations.

Loading images

set_image accepts a numpy array or a BioImage object:

# NumPy array (2D, CYX, CZYX, or TCZYX)
viewer.set_image(np.array(...))

# BioImage — lazy loading, full 5D support
from bioio import BioImage
import bioio_tifffile, bioio_ome_zarr

viewer.set_image(BioImage("image.tif",  reader=bioio_tifffile.Reader))
viewer.set_image(BioImage("image.zarr", reader=bioio_ome_zarr.Reader))

When a multi-dimensional image is loaded, sliders for T and Z appear automatically, and the Layers panel shows per-channel controls.