TechLila has been publishing since 2012. An archive that old creates an obligation most sites never face: a guide written for one version of macOS is still on the internet three versions later, still ranking, still being followed by someone who has no way of knowing when it was last true. Our fact-checking is built around that problem first and around new articles second.
What we check, and what “still true” means here
A claim on TechLila is anything a reader could act on at their keyboard. Three kinds carry different risks, so we treat them differently.
Instructions are the highest risk, because a reader follows them rather than reads them. Every menu path, keyboard shortcut, terminal command, and settings toggle is executed on the software version named in the article before it publishes. If we cannot reproduce a step, the step does not appear.
Numbers are prices, subscription tiers, storage limits, benchmark results, app counts, and market-share figures. Each traces to a named source and is stamped with the date we read it. Prices in particular are treated as perishable: a figure without a date beside it is a defect, not a detail.
Judgments are which app is worth paying for, which workflow is faster, which trade-off matters. These are ours, and we label them as ours. We do not dress an opinion as a measurement, and we say what a recommendation is based on.
Where our facts come from
We prefer the party that created the thing being described.
Vendor documentation, release notes, developer docs, and official support pages settle questions about how software behaves. Pricing comes from the vendor’s own pricing page, never from a roundup. Platform and hardware claims come from the maker’s technical specifications. Standards questions go to the body that publishes the standard.
When another publication reports something first, we credit them by name and link to their story, then find the underlying document and cite that too. A link to a story about a document is not a citation of the document.
We do not cite spec-aggregator sites, affiliate roundups that tested nothing, press releases with no data behind them, or AI-generated summaries. These recycle other people’s errors efficiently, which is precisely the failure mode we are trying to avoid.
How an article gets verified
Verification is a sequence of gates. An article does not advance until it clears the one it is sitting in.
Sources are captured before the writing starts. We fetch the page, keep the passage that supports the claim in its own words, and record the URL and retrieval date. Later paraphrase is checked against that captured passage, not against a memory of it.
Every claim is tagged to its source inside the draft. An untagged factual sentence is a build error. This is mechanical and it is unforgiving, which is the point.
A deterministic pass checks what machines check better than people. Numbers must match their source exactly, arithmetic must resolve, quotes must be verbatim, and dates must be internally consistent.
A semantic pass asks whether the source actually supports the claim. Matching numbers are not enough. A source that reports a figure for one region does not support a worldwide claim, and a projection does not support a statement of fact.
A named reviewer reads the finished piece. TechLila carries reviewer attribution on the article itself, and the reviewer is never the person who wrote it.
The verification record is archived against the article so a later editor can see which source supported which sentence.
When a claim fails
There are three ways out and no fourth. Re-source it from something better, narrow the claim until the source genuinely supports it, or delete it. Publishing a claim that failed verification is not on the list, and neither is softening the wording until the failure stops being visible.
Keeping the archive honest
Every article carries the date it was last reviewed, which is the date a person checked it, not the date a plugin touched the file.
Guides are re-tested when the software they describe changes in a way that moves a step. Statistics articles are rebuilt from current data on a schedule rather than when someone notices. When an article can no longer be made correct, we retire it instead of leaving it up with a fresh date on top.
Independence
TechLila earns revenue from advertising, sponsored placements, and affiliate links. Sponsored work is labeled and produced separately from editorial and is never published as editorial.
An affiliate link never changes a verdict, an ordering, or a score. Where we recommend a paid product, the reason is stated in the article and the reader can check it against the sources we cite. Advertisers see no coverage before it publishes and have no input into it.
Corrections
If a figure is wrong, a step no longer works, or a source has moved, tell us through the contact page. We reply to every report, correct the article, note what changed, and update the verification record behind it.
Last reviewed: 29 August 2026.