Spyder-notebook brings Jupyter notebooks into the Spyder IDE

What the plugin actually does

Spyder-notebook is a Spyder plugin to use Jupyter notebooks inside Spyder, and the README is careful about the scope. Currently it supports basic functionality such as creating new notebooks, opening any notebook in your filesystem and saving notebooks at any location. That is a modest feature set and it is worth being clear about it, because the plugin is not a full notebook environment with cells, execution and inline plots. It is a bridge that lets notebooks live inside the IDE's project and file system, which means the notebook gets the same file switcher, the same project tree and the same save behaviour as the rest of your code. The plugin is published to PyPI as spyder-notebook, to conda-forge, and it is translated through Crowdin, with the badge in the README pointing at the translation project. The most recent releases are v0.7.0 in October 2025, v0.6.2 in April 2025 and v0.6.1 in September 2024, and the last push was on 2026-07-20.

The two integrations that make it more than a file opener

Two features lift the plugin above simply viewing notebook files. The first is the file switcher, and the README says you can use Spyder's own file switcher to easily switch between notebooks, which means the plugin hooks into a control that already exists in the IDE rather than adding a parallel navigation scheme. The second is more interesting. You can open an IPython console connected to the kernel of a notebook to inspect its variables in the Variable Explorer. That is the feature that ties the plugin to Spyder as a whole, because the Variable Explorer is one of the IDE's main panels and it is where a Spyder user expects to look at the state of a program. With the console attached to the kernel, the variables in that kernel show up in the same panel as the variables of the running script, so you can inspect a notebook's state without leaving the IDE or opening a second application. That is the workflow the plugin exists to support.

Install it with conda, because pip does not work well

The installation section is short and has a clear recommendation. The plugin can be installed using the conda package manager as follows:

conda install spyder-notebook -c conda-forge

The plugin is also available on PyPI, and the README says that in principle you can use pip to install it, but that this often leads to installations that do not work properly. The wording is a candid admission rather than a hard prohibition, and the practical advice is to use conda-forge and treat the PyPI route as a fallback. There is a second limitation that is stated as a note and it is a hard one. At the moment it is not possible to use this plugin with Spyder's stand-alone installers, and the project says it is working to make that a reality in the future. Since Spyder's own distribution channels include a standalone installer that many users choose, that note excludes a large share of the Spyder user base and is the first thing to check before anything else.

The dependencies are minimal and the kernel is the hard part

The project depends on Spyder and on Notebook from Jupyter, and that short list is the whole dependency story. The plugin builds on the Jupyter Notebook package rather than on a separate notebook server implementation, which means the kernel you connect to is a kernel you can start from the same environment Spyder is running in. That is what makes the IPython console attachment work, since a console connected to a kernel has to be talking the same protocol the notebook server expects. The practical constraint that follows is that the kernel has to be startable, which means its dependencies have to be installed in the active environment, and the plugin's own dependency on the Notebook package means the Jupyter stack is present by definition. The tests are run with pytest, the repository has a pytest.ini and a conftest.py at the root, and coverage is tracked through codecov with a .coveragerc, while the test workflow badge points at a GitHub Actions run-tests workflow.

The setup script shows the plugin ships a JavaScript server

The setup.py in the repository explains more about the architecture than the README does. It is a namespace package project, since it uses find_namespace_packages, and it reads the version from the module's _version.py file by parsing a VERSION_INFO tuple out of it with ast.literal_eval, which is a common pattern for keeping the version in one place. The interesting part is the custom sdist command, a class extending setuptools' sdist whose run method compiles JavaScript and then does the default work. It checks whether a build directory inside the server directory exists, and if it does not, it runs jlpm install and jlpm build in the server directory. The build directory is defined as the static folder inside the plugin's server directory, so the plugin bundles a built JavaScript front end rather than running a separate notebook server from source. The script also echoes each command before running it, and it handles Windows differently, using list2cmdline there and shlex quoting elsewhere, which is the sort of detail that only matters when someone tries to build a source distribution on Windows.

Repository layout and the people behind it

The top level of the repository shows a project with a full support structure rather than a small add-on. There is a spyder_notebook/ package directory, a doc/ directory, a requirements/ directory, a setup.py, a conftest.py, a pytest.ini, a codecov.yml, a crowdin.yml for the translation pipeline, a MANIFEST.in, a CHANGELOG.md, a CONTRIBUTING.md, a RELEASE.md, a .ciocheck for the open collective, and a .coveragerc. The funding section is more substantial than most plugin READMEs. Spyder and its subprojects are funded through the Chan Zuckerberg Initiative and NumFocus, along with donations from users around the world collected through Open Collective, and the README carries the badges for all three along with the sponsors list. The development section points at the Contributing Guide for how to set up a development environment, and the contact route is the Spyder group on Google Groups or an issue in the repository. The licence is a custom one that GitHub cannot classify, with the file at the root, and the setup script header describes the code as being under the MIT License terms, so the terms are worth reading directly rather than inferring from the badge.

Editorial conclusion

Adopt Spyder-notebook if you already use Spyder and want to work in Jupyter notebooks without leaving the IDE, particularly if you rely on the Variable Explorer to inspect a kernel's state. Do not adopt it if you install Spyder from a stand-alone installer, since the README says the plugin cannot be used that way yet, or if you need a package manager other than conda, since pip installs are reported to often lead to installations that do not work. Verify first that your Spyder comes from conda, that the Jupyter Notebook package is present, and that the kernel you connect to is one you can start inside your environment.

Frequently asked questions

Official sources

1. Issues
2. README
3. Releases
4. spyder-ide/spyder-notebook on GitHub