How we evaluate
Every capability and price in this publication is read from the vendor's own documentation and carries the date it was checked. These are the rules that produce it.
Primary sources only
Every price, limit and feature gate is read from the vendor's own pricing page or documentation, and linked to the page it came from so you can follow it.
Every claim carries its date
Bot pricing and feature gates move. Each fact is stamped with the day it was read from the source, so the currency of every number is visible on the page.
List prices, not promotional prices
Where a vendor is running a sale, we record the list price and note the promotion separately. A discount that expires next month is not what the product costs.
Unverified is not a verdict
When a vendor's site does not expose something we need — several are client-rendered applications that show nothing to a reader without scripting — the table says Not verified. It never shows a guess styled as a finding.
Rankings state their criteria
Every verdict names the criteria it was reached on, and what it is not weighted by. Where a specialist leads its own category, we name it as the specialist for that need — a ranking is only useful if you can see the reasoning behind it.
No invented authority
No fabricated ratings, no unnamed testimonials, no invented author credentials or years of experience. Articles are published under the desk that wrote them.
Corrections
We correct in place and note what changed. If a price, limit or feature has changed since the date shown, tell us and we will re-check it against the vendor's page and update the date. Corrections from the bot developers themselves are welcome and get the same treatment as any other.
Disclosure. Articles are written in-house with AI assistance in drafting. Every factual claim is checked against a primary source before publication. We do not attach invented author names or fictional credentials to anything we publish.