Most AI explainers are either a paywalled course or a blog post that was true six months ago. This is ~700 free pages instead: 250+ glossary terms, 45 model pages, 65+ tool reviews, use-case guides and a news blog. No signup, no paywall, no email gate. Benchmark numbers always name their source and the date they were checked. Makers can also add their own AI tool to the catalog for free at /submit-ai-tool.
Hi Product Hunt 👋
I built HowAIWorks.ai because most AI explainers are either a paywalled course
or a blog post that was true six months ago. This is neither: ~700 free pages —
250+ glossary terms, 45 model pages, 65+ tool reviews, use-case guides and a
news blog. No signup, no paywall, no email gate.
Two things I care about:
1. Editorial rules that survive contact with a fast-moving field. Benchmark
numbers name the benchmark, the source and the date they were checked.
An Artificial Analysis score is never quoted without its index version,
because versions aren't comparable. Pages describe mechanisms rather than
this month's leader.
2. Every page ships a plain-Markdown twin at the same URL + .md — just the
article, no navigation, no scripts. There's an llms.txt too. If you point
an agent at a page, it gets clean text instead of parsing page chrome.
Try it: https://howaiworks.ai/glossary/r...
For makers here specifically: you can submit your own AI tool to the catalog
for free at https://howaiworks.ai/submit-ai-... — free listings get an ordinary
followed link, and only paid placements are marked sponsored.
I'd genuinely like to hear where the writing is thin or wrong. Tell me which
term or model page you'd check first, and I'll fix what you find.