You’ve encountered a problem? See if you can find it in below table. If not, please get in touch via issues on Github.




Installing navis for Blender fails with “Python.h” not found.

Manually download and copy required files into Blender’s Python directory. Follow the instructions here.


Trying to use 3D plots raises an error referencing VisPy backends

Choice of VisPy backend depends on your python distribution, OS, hardware, and how you run it. A minimal navis installation chooses no backend, but all backends <> are available as navis extras; e.g.

pip install ‘navis[vispy-pyside6]’

You may need to experiment to find the best backend for your environment.

3D plots with VisPy as backend use only one quarter of the canvas.

Try installing the developer version from GitHub ( As one-liner:

git clone && cd vispy && python install --user

3D plots using Plotly are too small and all I can see is a chunk of legend.

Sometimes plotly does not scale the plot correctly. The solution is to play around with the width parameter:

fig = navis.plot3d(neurons, backend='plotly', width=1200)

Opening a vispy 3D viewer, throws a long list of errors ending with something like this:

RuntimeError: Error while fetching file
Dataset fetching aborted (<urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed (_ssl.c:847)>)

For reasons beyond me, vispy does not include the font to render text so it has to download it on first use. If this fails with an SSL error, do the following once:

import navis
import ssl
ssl._create_default_https_context = ssl._create_unverified_context
v = navis.Viewer()

This temporarily disables SSL verification to allow download of the font. I recommend restarting the Python session afterwards!


Instead of a progress bar, I get some odd message (e.g. Hbox(children=...) when using navis in a Jupyter notebook.

You probably have ipywidgets not installed or not configured properly. One work-around is to force navis to use standard progress bars using navis.set_pbars():



Navis starts viewers for plots when I don’t want it to.

Set the environment variable NAVIS_HEADLESS to "True" before navis is first imported to disable viewers (good for use on servers).

Navis interferes with my logging configuration.

By default, navis configures sensible defaults for logging (helpful for jupyter notebooks, scripting, and exploratory REPL use). Set the environment variable NAVIS_SKIP_LOG_SETUP to "True" before navis is first imported to disable this (good when using navis as a library). Use navis.config.default_logging() to manually run the log setup.