Skip to main content

ComfyUI ModuleNotFoundError / "No module named" — custom node won't import

A ModuleNotFoundError at ComfyUI startup almost always means a custom node's Python dependency didn't install — the node folder is there, but a package it imports is missing.

Also seen as: "ModuleNotFoundError", "No module named", "ImportError", "cannot import"

Direct answer

How do I fix "Missing Python Package"?

A ModuleNotFoundError at ComfyUI startup almost always means a custom node's Python dependency didn't install — the node folder is there, but a package it imports is missing.

Key facts

  • iA custom node's requirements.txt never ran (headless installs skip it).
  • iA from-source build (sam2, llama-cpp-python, insightface) failed or timed out.
  • iPython/torch version mismatch — a wheel built for a different version.

Next step

  1. 1Reinstall the node's requirements: pip install -r custom_nodes/<node>/requirements.txt.
  2. 2For from-source packages, use --prefer-binary and a longer build timeout, or a prebuilt wheel.
  3. 3Paste the workflow into a checker that resolves each node to its repo + install command before you deploy.
  4. 4Match torch/CUDA versions to what the node was built against.

Not a fit when

  • Reinstalling the node folder won't help if the missing package is a system build dependency.

Not sure this workflow will run before you rent a GPU?

Paste your ComfyUI workflow JSON into the free checker — it flags missing nodes, model requirements, and the VRAM/GPU you actually need. No signup.

Check my workflow freeSee GPU pricing

Other ComfyUI errors

Fix

GPU Memory Exhausted

Fix

Model Not Found

Fix

Network Connection Failed

Fix

GPU Driver Error

Fix

Process Terminated

Fix

Permission Denied

Frequently Asked Questions

What causes "Missing Python Package" in ComfyUI?

A custom node's requirements.txt never ran (headless installs skip it).

How do I fix "Missing Python Package"?

Reinstall the node's requirements: pip install -r custom_nodes/<node>/requirements.txt. For from-source packages, use --prefer-binary and a longer build timeout, or a prebuilt wheel.