Jordan Hale
Editor and lead reviewer, Meln. Twenty years in software, the last six on AI products, including the team that once shipped a product on this very domain. Now reviewing the tools instead of selling them.
Background
Jordan started out as a backend engineer in the mid-2000s, the era when "machine learning" meant a logistic regression and a very patient DBA. Over the following fifteen years the job titles drifted from engineer to product lead to, eventually, the person in the room who had to decide what the pricing page said. That included two B2B software companies, one of which was acquired, and a long stretch running product for a text-analytics team whose detector, in hindsight, was nowhere near as good as the sales deck claimed.
From 2020 onwards the work was all AI: a writing-assistant startup, a consultancy helping mid-size companies decide which language-model vendor to trust, and finally the small team that built an earlier product on the meln.ai domain. That last company wound down in 2025. Jordan kept the domain, and the habit of reading pricing pages with a calculator open, and started writing the reviews he wished had existed when he was the one buying.
Why a reviewer and not a builder
Partly fatigue, partly conviction. After six years of watching launch decks quote the one benchmark that flattered the model, Jordan came to think the useful job was not building the next tool but telling people plainly what the existing ones do, what they cost once the "per word" footnote is applied, and which claims survive contact with an independent test. There are more than enough people building. There are very few writing honestly about billing.
How Jordan reviews
Pricing first, read from the vendor's page on a stated date and converted into what a realistic user would actually pay in a month. Then evidence: independent benchmarks, regulator actions, court records and published third-party tests, each attributed, with vendor claims labelled as claims. Then the complaints, read by theme across review aggregators, because a product's billing behaviour is part of the product. Then a plain verdict including who should skip it. Jordan does not claim to have run tests that were not run, and does not quote a number that cannot be traced to a source. The full method is on our how we test page.
Categories covered
AI humanizers, AI detectors, AI companion apps and AI trading bots, plus standalone reviews of the tools in each. These are the four corners of the AI market where the gap between marketing and reality is widest and the refund policies are least forgiving, which is exactly why they got covered first.
Off the clock
Jordan lives in the Pacific Northwest, cycles badly, reads court filings for fun, and still maintains a spreadsheet of every software subscription he has ever cancelled and why. Several of the "who should skip it" paragraphs on this site started life as a row in that spreadsheet.
Disclosures
Jordan holds no equity in any company reviewed on Meln and receives no payment from vendors. Meln as a publication earns affiliate commissions, described on our disclosure page. Jordan does not trade with any of the bot platforms reviewed and does not hold accounts with the companion apps beyond what is needed to verify a feature. Corrections are welcome via the contact page and are logged on the page concerned.