Install the Jupyter workbench locally¶
If you already have a hosted IDE account, enter the platform without installing it again. This page is for readers who want to run notebooks on their own computer.
Obtain a matching pair of packages¶
Get the SDK and CASCAQit-Jupyter wheels, checksums and compatibility notes from the maintainer. Private repository releases require access permission. A wheel includes the Python companion and the prebuilt frontend extension, so installing it does not require Node.js. Building the frontend from source does.
A successful package installation does not prove compatibility. The standalone IDE lock specifies SDK 1.0.8a0, cascaqit-jupyter 0.2.0a1+ide.ee39d64.1 and Bokeh 3.4.3. That companion includes adaptations for the SDK operation API and is not identical to upstream 0.2.0a1. These versions identify that distribution; they are not a claim about every live environment.
Do not install the learning site's build dependency lock into Jupyter. The companion requires its Python Bokeh and bundled BokehJS to match; documentation rendering uses a separate environment.
Use a dedicated virtual environment¶
The following example is for macOS/Linux. Place the two supplied wheels in wheelhouse and replace the uppercase filenames with their actual names. Do not substitute unverified packages with similar names.
python3.11 -m venv .venv
source .venv/bin/activate
python -m pip install ./wheelhouse/SDK_WHEEL.whl ./wheelhouse/JUPYTER_WHEEL.whl \
"jupyterlab>=4,<5" "notebook>=7,<8" ipykernel
python -m pip check
jupyter labextension list
jupyter lab
In Windows PowerShell, use py -3.11 -m venv .venv, followed by .venv\Scripts\python.exe and .venv\Scripts\jupyter.exe for the corresponding commands. Check that @cascaqit/jupyter is enabled and reports OK. Then verify the active kernel using the notebook version check.
Use the local URL and access token printed by Jupyter. Keep authentication enabled. A notebook executes ordinary Python; read code from other people before running it.
Code, completion and Git¶
Notebooks, quantum editors and result displays work without code-server. For the Code view, follow the Jupyter project's setup instructions for the ide extra and code-server. Its installer supports macOS/Linux. Windows users can use notebooks; the script does not establish native Windows support for code-server.
The ide extra also provides Python language services and a Git panel. Use Ctrl+Space or Jupyter's Tab completion; debugging depends on the installed Jupyter and ipykernel. Select the same Python environment in Code and Notebook, while keeping in mind that they do not share variables in memory.
GitHub Codespaces¶
Upstream offers a Codespaces entry for accounts with read access to the private Core repository. Its configuration downloads a fixed SDK release wheel, rather than following this site's dev version. Open JupyterLab from the PORTS panel and keep the port private. Run jupyter server list in your own Codespace terminal if you need its access token.
A Codespace is a separate development environment, not an account on the standalone IDE platform. Check the current repository instructions, permissions and available allowance before using it.
Continue with Your first notebook experiment. If dependencies conflict, retain the installation log and version list and consult troubleshooting.