Click Coded

People are starting not to trust AI, and for any business that uses it that’s a slow-motion crisis. Click Coded is an openly AI-operated studio attacking that problem from the only angle that isn’t self-serving: make every claim checkable, starting with its own. When machines make claims free, checkable is the only currency left.

It’s fronted by Alexander K. Eliot. A disclosed AI persona, a pen name, marked as AI-operated on every single page, mentored by me. The disclosure isn’t a disclaimer at the bottom. It’s the product thesis.

What it actually ships:

• A free AI Visibility Checker: paste a URL, get a score out of 100 on twelve technical checks in thirty seconds, no signup. Full rubric published.
• The Checkable Standard: an open spec for making an AI-operated business’s claims auditable by anyone. Click Coded wrote it and lives under it first.
• The receipts: 39 published audits naming names. Of 81 crisis hotlines, only 4 tell AI assistants they exist. Of 41 major newsrooms, 34 block AI crawlers on purpose. Mayo Clinic and Johns Hopkins return a hard 403 to the assistants patients ask health questions.
• Compute grants: the same trust work, tithed free to crisis services and nonprofits.

The move that makes it real: it publishes its own failures. Its Gumroad storefront scored 45/100. Dead last on its own twenty-site benchmark. It said so on the product page. This homepage scored 42/100 against its own checker before the redesign, and the before-and-after is published as a case study.

A benchmark that hides its own bad numbers isn’t a benchmark. That’s the whole company.