Your pipeline should not open every upload just to learn it was never video.
You get a batch of paths. Some are reports. Some are exports someone renamed. Before you spin up probes or transcodes, you need a cheap answer: does this name look like video at all?
Video Extensions is that first gate. One import, and you can treat bare extensions and full paths the same way.
Know what looks like video before the heavy work starts.
What you get
- A quick yes or no on extensions and paths so upload validators and ingest jobs can reject obvious mismatches early.
- Thirty-six known video extensions in one place, including common containers, streaming types, professional formats, and legacy names you forget to add by hand.
- Case-insensitive, dot-aware matching so .MOV, movie.MP4, and bare mkv strings behave the way you expect.
- Dotfile-style names handled when the whole filename is effectively the extension, such as a lone .mov path.
- A shared list you can reuse when you need to iterate allowed types or layer your own rules on top.
- No extra packages beyond what you already ship, so the check stays small and predictable.
Picture your Monday
Your ingest job watches a folder of uploads. For each filename, you ask whether it looks like video before you queue anything expensive. The obvious text and image files drop out in milliseconds. The rest move on to deeper checks you already trust. An illustrative case, but the pattern is the same in APIs, desktop tools, and batch rename scripts.
Ready to try it?
- See it on GitHub for usage, tests, and how it mirrors the widely used video-extensions dataset
- Get it on PyPI






