How we cover the AI tooling space.

How we pick tools to cover

We track tools that meet at least two of these criteria: real funding or real revenue (not pre-launch concept), real production usage at companies we can point to, or meaningful technical differentiation we can independently verify against the vendor's docs and behavior. Pre-launch tools and rebadged generic-LLM wrappers don't make the cut.

Once a tool is in the index, it stays — even if it loses momentum. Hiding declining tools would distort the historical picture readers use to judge category trajectories. We update each tool page as the trajectory becomes clearer.

What sits behind each review

Our reviews draw on three sources, weighted in the order they appear here:

• Hands-on use, where we have it. Some tools we have used directly — daily or for specific projects. Where this is the case, the review includes observation that only emerges from real use (failure modes, hidden friction, what works at scale that doesn't show in demos).
• Structured operator reports. For tools we haven't deployed at production scale ourselves, we draw on conversations with operators who have, plus public field reports, postmortems, and benchmark data. We say "operators report" or similar framing when that's the basis.
• Vendor documentation and direct testing. We read the docs, sign up for trials, test the product against the use cases the marketing pages claim, and verify pricing against the current pricing page (not a screenshot from six months ago).

We do not pretend to have used every tool we cover at production scale. That would be a fabrication, and the editorial value of the site depends on not fabricating. Where the review draws on synthesis rather than first-hand use, the framing reflects that.

What "verified" means on a tool page

Every tool page carries a Verified [Month Year] stamp. That means, on or around that date, we re-read the vendor's pricing page, re-verified the feature claims in the review against the current product behavior or current docs, and updated anything that had changed. The stamp is the freshness signal, not a substitute for "we just retested everything."

We aim to re-verify each tool quarterly. Pricing or feature claims older than six months get flagged on the page until refreshed.

How we structure reviews

Every tool review follows the same skeleton:

• What it is. A description we wrote, not the vendor's marketing copy.
• Where it shines. Specific strengths grounded in the product's actual behavior or documented capability.
• Where it falls down. Honest weaknesses. If we can't list at least two, the review isn't ready to ship.
• Notes on using it. Synthesized observations from our use and from operator reports — what's different from the marketing pitch, what to look out for, what the failure modes are in practice.
• Pricing and hosting. What you'll pay, where the product runs.

How comparisons work

Head-to-head comparisons (/compare/) and alternatives lists (/alternatives/) follow the same standards as tool pages. We pick combinations with real search and decision intent — not the full Cartesian product — and we write a verdict that says who should pick which. The verdict is opinion, not benchmark — we say so on the page.

Where we have direct experience with the tools in the comparison, we say so. Where the comparison is built from synthesis of public sources and operator reports, the language reflects that. We avoid the common pattern of asserting a side-by-side experimental result when no such experiment was run.

How we make money

Some tools we cover have affiliate programs. When you click a link to a tool from our site, we may earn a fee if you sign up. Several things we promise:

• Affiliate status never changes the editorial. Tools without affiliate programs get the same coverage as tools with them. If a tool with an affiliate program is bad, we say so.
• We don't accept paid placements. No "sponsored review" content masquerading as editorial.
• We don't accept money to soften coverage. A vendor cannot pay to remove a "where it falls down" section.

If you'd rather not contribute affiliate revenue, navigate to the vendor directly — our editorial doesn't change either way.

Conflicts of interest

When the editorial team has a financial relationship with a tool we cover (advisor, investor, employee, family member, etc.), we disclose it on that tool's page. We don't accept stock or equity in exchange for coverage.

How we use AI to write about AI

We use AI tools throughout our editorial process — drafting outlines, doing research, polishing copy, generating structured data. We don't pretend otherwise. What we won't do: ship an AI-generated article without a human editor running through it line by line, fact-checking the claims, and adding the opinion only a human who's reasoned through the trade-offs can add. That last layer is the entire point of the site.

Corrections

See something wrong? Email fix@airunsmycompany.com and we'll update the page or push back if we disagree. Substantive corrections are logged at the bottom of affected pages with the date and what changed.

Coverage requests

If you're a vendor pitching a tool: /contact has the details. We don't guarantee coverage, and we don't trade coverage for sponsorships.