Loading…
Loading…
About
I'm Brandon Perkins — a software developer for the last 14 years, an arrowhead enthusiast for basically my whole life, and the person who builds and runs ArtifactVerify PointCheck.
To be clear about who I am: I'm not a professional authenticator, and I'm not an archaeologist or lithics expert. I'm a collector who has spent decades reading typology guides and walking creek beds. By day I'm a Software Engineer. There are genuinely skilled authenticators in this hobby, with eyes trained by decades of handling artifacts, and their work matters. But there aren't enough of them to go around, and buying points online still too often comes down to a leap of faith.
That frustration is shared across the community, and many collectors have simply accepted it as the cost of doing business. What struck me, coming from software, is that the gap isn't expertise. It's measurement. Even the most respected authentication is delivered as an expert judgment with no published error rate attached; not because anyone is hiding something, but because the hobby has never had the infrastructure to measure one. And if something isn't measured, it's hard to trust at scale - and nearly impossible to improve. And so I set out to build the tool I kept wishing existed.
PointCheck is my attempt to close this gap. The system matches your point's shape against a reference corpus of over 1,000 North American types, runs a visual authenticity assessment, and returns a probability, not a verdict. And we publish exactly how often it's right: the measured accuracy numbers are on the methodology page, including the ones that aren't flattering. When the model can't verify something like orientation, background removal, a type match - the report says so.
I encourage you to check the numbers, read the evidence in your results, cross-reference with well known arrowhead sources online, and - for anything high-stakes - get a hands-on examination from a professional authenticator on top of the screening. I would love to hear your feedback as there is always room for improvement.
The methodology page walks through the full pipeline — shape matching, the AI assessment, and the measured accuracy of both.
Read the methodology