Check your Pavatar
A website publishes a Pavatar in one of three ways. You can check all three in under a minute with curl — these are the same steps the original “Check your Pavatar!” script followed.
Replace https://example.com/ with the website you want to check. Stop at the first step that finds something — that is the order of precedence in version 0.3.0, and the diagram on Pavatar explained replays the same three steps.
1. HTTP header
curl -sI https://example.com/ | grep -i '^x-pavatar:'
A line like X-Pavatar: https://example.com/me.png means the site publishes a Pavatar at that address. X-Pavatar: none means the owner refuses avatars — stop here.
2. Link element
curl -s https://example.com/ | grep -io '<link[^>]*rel=["'"'"']pavatar["'"'"'][^>]*>'
Look for <link rel="pavatar" href="…">. The href must be an absolute URL, or none.
3. pavatar.png
curl -sI https://example.com/pavatar.png | head -n 1
A 2xx status means the file exists. For a page deeper in the site, the specification first checks pavatar.png in the page's own directory, then at the site root.
Is the image valid?
A valid Pavatar is an 80×80 pixel GIF, PNG or JPEG of at most 409,600 bits (51,200 bytes), served with the matching Content-Type:
curl -sI https://example.com/me.png | grep -i -E '^(content-type|content-length):'
The history of the protocol is on the Pavatar archive page.