python-notebooks-async

To run asyncio in Jupyter notebooks without RuntimeError, use top-level await directly in cells. Notebook kernels already own an active event loop, so avoid creating new loops with asyncio.run to prevent conflicts.

Yes, you can use asyncio.gather for concurrent tasks and TaskGroup for structured concurrency in IPython. These patterns organize asynchronous execution safely within the notebook kernel's host-owned event loop.

asyncio.run fails in VS Code notebooks because the kernel already operates a host-owned event loop. Calling asyncio.run attempts to create a conflicting new loop, triggering a RuntimeError instead of executing your async code.

The best way to manage top-level await in notebook environments is treating the kernel as a loop-owned host. Prefer top-level await in cells and keep reusable async logic in regular Python modules rather than patching global loops.

You should not use global event loop patches for async notebook workflows when the kernel already manages the loop. Minimizing global patches and utilizing top-level await ensures reliable execution without unpredictable behavior.