Python import failed

Troubleshoot python environment errors using the traceback and your ComfyUI environment.

Find the first failing import

A red or unavailable custom node can originate from an earlier startup exception. Read the startup log from the beginning of the affected package's import. The final missing-node dialog is often a consequence rather than the root cause. Record the exact package path, exception name and missing module.

Use the environment that runs ComfyUI

  1. Record how ComfyUI is launched: desktop, portable bundle, a virtual environment or a container.
  2. Identify the Python interpreter belonging to that installation. Installing a package into an unrelated system Python will not resolve an import in the ComfyUI environment.
  3. Open the maintained custom-node repository or registry entry and read its supported Python, PyTorch and platform requirements.
  4. Follow those instructions for the correct environment. Review the package and changes before installation; do not copy an arbitrary installation command from a log.
  5. Restart through the supported launcher and confirm that the package imports successfully before reloading the original workflow.

Separate missing packages from version conflicts

ModuleNotFoundError usually points to a module that cannot be found. An ImportError involving an attribute or symbol can instead mean a incompatible version or binary build. In that case installing more packages blindly may make the environment harder to recover. Keep a record of installed versions and use a backup or separate environment when testing upgrades.

When not to change the environment

An optional acceleration extension may fail while a supported fallback still runs. Confirm whether startup completes and a small control workflow succeeds. A warning is not proof that every graph is broken. OnBeg lists the evidence and class names; installation status requires your running environment.

Continue with local diagnostics

Check the workflow · Inspect the log

References: ComfyUI troubleshooting · Custom-node installation.