Every browser still claims to be Mozilla/5.0 for reasons that stopped mattering twenty years ago. Analytics code has to untangle that, and testing it needs agents from every platform — including the bots, which look nothing like the rest.
What you can control
- Desktop, mobile and bot agents, each with the correct platform token.
- Parsed browser, version and OS columns alongside the raw string, so you can check your parser against them.
- Bot agents include the real crawler strings plus
curlandpython-requests, which is what most scraping traffic actually sends. - Optional Sec-CH-UA client hints, the header set that is gradually replacing the User-Agent string.
Testing ideas
- Device detection. Feed desktop and mobile strings to your detection code and check each lands on the right layout.
- Bot filtering. Include crawlers to make sure analytics exclude them and your robots handling treats them correctly.
- Client hints. Turn on
Sec-CH-UAheaders; Chromium browsers increasingly report detail through client hints instead of the user-agent string. - Parser drift. Compare the parsed browser and OS columns with what your own parser reports.
What this is not
Browser versions here are a fixed recent set and will age. Do not use them to test version-gating logic that depends on the current release.