Data sources
Four public sources, always kept apart, so you can judge how strong a claim is.
Subprocessor lists
The company names the vendor itself, usually for legal reasons. This reveals backend tools nothing else can see.
Privacy policies
The vendor is named in the policy, with a role when one is clear. Wider coverage, but weaker than a list.
Website code
Loaded by the website, set in its DNS or running its job board. We keep the exact signal we saw.
Job postings
Asked for in open postings. It shows what a team works with, not that the company has bought it.
How a claim gets its level
Each record is marked confirmed, likely or possible, from the kinds of evidence behind it and how recently we saw them. Agreeing sources raise a claim, and old evidence fades.
What we promise
- A name we publish has to appear in the source text.
- When we cannot validate a page, we show nothing rather than guess.
- "We found nothing" never means "they use nothing". Pages behind a request or a login stay unread, and we say so.
- Personal data is removed from job postings before we store them.
We read only public pages. How our crawler behaves.