Fact · Voice privacy and security
When you load a named Whisper model, the openai-whisper loader fetches the model file from its download address with urllib, unless a file with the right checksum is already in the cache folder.
Checked against its source on October 17, 2026
In the source's own words
“with urllib.request.urlopen(url) as source, open(download_target, "wb") as output:”
openai/whisper: whisper/__init__.py (model download and loading code), OpenAI, GitHub (openai/whisper, main branch), 2026, _download function, the download step after the cache check.
Open the original- What it measures
- The loader first checks whether the model file already exists in the download folder and that its SHA256 checksum matches. Only when the file is missing or does not match does the code open a network connection and save the file to disk.
- What it does not show
- This is one line of the loader's code, read on the main branch of the project, not a documented guarantee. Other Whisper apps and forks download models in their own ways, and the code can change in later versions.