Editorial standards
Last updated 26 October 2026
Most tool sites describe what their tools can do. The pages here also describe what they cannot, where your file goes, and when you would be better off using something else. That is a deliberate choice and this page explains how it works.
The short version
- Every tool page is written from that tool's own source code, not from competitors' pages.
- If a page says your file stays in the browser, someone read the route to confirm it.
- No figure appears unless it can be verified. No ratings, no user counts, no invented benchmarks.
- Every page states the tool's limits, including the cases where you should use something else.
- Advertising has no influence on any page, and there are no affiliate links.
How a tool page gets written
The starting point is the code, which is an unusual way to write this kind of page and the reason these read differently from most.
Before a page is written, the route behind the tool is read: what it accepts, what it does with it, whether it makes a network request, and where it gives up. That answers the questions readers actually have, and several of the answers were surprising enough to change what the page says.
The JPG to JPEG converter does not convert anything, because the two are the same format and the code copies the bytes unchanged — so the page describes a rename rather than a conversion. The PDF editor flattens every page to an image when it saves, which means the result has no selectable text, so that is the first thing its page tells you. Neither fact is flattering and both matter more than anything a marketing description would have said.
Writing from competitors' pages instead is how sites end up repeating claims nobody checked. It is also how this site previously came to state that a tool kept your files local when it uploaded them.
What gets checked before a page is published
- Where the data goes. Confirmed by reading the route. Several tools are hybrids that work locally below a size threshold and upload above it; those thresholds are stated because the difference matters.
- How long anything is kept. Taken from the actual job expiry and cleanup interval, which work out to about 90 minutes for most server tools and seven days for a protected PDF share link.
- Every number. Thresholds, size caps and supported formats come from the code. Anything that cannot be traced to a source is cut.
- That the page matches its own structured data. Questions described to search engines have to appear on the page, which is a requirement that this site previously broke on 75 pages.
- The awkward cases. The tool itself is tested the way the About page describes — wrong extensions, truncated files, empty files, very large files, and the network panel left open to confirm a local tool really is local.
What never gets published
These are rules rather than preferences, and each exists because the site previously broke it.
- Invented figures. No compression ratios, processing speeds, accuracy percentages or user counts. Where a figure would help and is not verifiable — a messaging app's compression settings, a country's passport photo specification, a rate of return — the page says so and points at the authority instead.
- Star ratings. Every tool page once carried a rating in its structured data, with values varying per tool and no review anywhere on the page. All 144 were removed and none replaced.
- Blanket privacy claims. “Your files never leave your device” was on every page including the ones that upload. Each page now speaks for its own tool.
- Hidden content. Nothing is described to a search engine that a visitor cannot read on the page.
- Padding. No page is lengthened to reach a word count. A short page that answers the question beats a long one that circles it.
On AI assistance
Drafts are produced with AI assistance. They are not published as drafted.
Every factual claim in a draft — what the tool does, where your file goes, what the limits are, which formats it accepts — is checked against that tool's source code before the page ships, and anything that could not be verified was removed rather than reworded into something vaguer. Several pages ended up saying something unflattering about their own tool precisely because that check contradicted the draft.
Saying this plainly seems better than the alternative. What matters for a reader is whether a page is accurate and useful, not how the first draft appeared, and a site claiming otherwise would be making a promise it could not evidence.
When something is wrong
Pages get things wrong. The useful question is what happens next.
Corrections are made directly, and significant ones are recorded rather than quietly edited away. The About page carries a section listing what this site previously got wrong, including the fabricated ratings, the false privacy claim in the footer, and the HEIC converters that uploaded photos to a server from pages promising the opposite. That list is not there for decoration — a trust page with nothing to admit is harder to believe than one that admits something.
If you find an inaccuracy, email hello@omniwebkit.com. Reports about a claim being wrong are dealt with ahead of feature requests, and the contact page explains what makes a report actionable.
Advertising and editorial independence
The site is funded by advertising served through Google. We do not choose the advertisers, we have no relationship with any of them, and no advertiser has ever requested a change to a page.
There are no affiliate links anywhere on this site. When a page recommends using a desktop application or a paid service instead of a tool here — which several do — there is nothing to be earned from it. That is the point: the recommendation is only worth reading if it costs us something to make.
What advertising is allowed to affect is which tools exist, since a tool nobody visits is not worth maintaining. What it is not allowed to affect is what a page says about the tool once it exists. The privacy policy sets out what Google's advertising cookies do and how to opt out of them.
Scope and what this does not cover
These standards apply to the 174 tool pages and to the About, Contact, Privacy and Terms pages. They are a description of current practice, not a contract, and they will be revised as practice changes.
Two limits worth naming. Nothing here is professional advice: the calculators explain their arithmetic and the health and finance pages state plainly that a clinician or a regulated adviser is the right place for a real decision. And no page promises an outcome — not a search ranking, not an approval, not a guarantee that a file will convert cleanly.
Frequently asked questions
Who writes the content on this site?
Deepak maintains the site alone. There is no editorial team, no freelance writers and no contributors. Where a page sounds like it was written by one person who read the code, that is because it was.
Do you use AI to write the tool pages?
Drafts are produced with AI assistance and are not published as drafted. Every claim about what a tool does, where your file goes and what its limits are is checked against that tool's own source code first. Anything that could not be verified was cut rather than softened.
How do you know where a tool sends my file?
By reading the route that handles it. Each tool page states whether the work happens in your browser or on a server, and several are hybrids with a size threshold, which is given. Those thresholds came from the code, not from a guess.
Why do some pages tell me not to use the tool?
Because sometimes that is the correct advice. The PDF editor flattens pages to images, the Word converter is less faithful than exporting from Word itself, and a long audio file is better handled by a desktop application. Omitting that would make the pages less useful, not more.
Do advertisers influence what the pages say?
No. Advertising is served by Google and we have no relationship with any advertiser, no sponsored placements and no affiliate links anywhere on the site. Nobody has ever asked for a change to a page, and there is no mechanism by which they could.
Why are there no star ratings or user counts?
Because we would have to invent them. The site previously displayed star ratings on every tool page and claimed 50,000 daily users, with no source for either. Those were removed and nothing replaced them.
What happens when a page turns out to be wrong?
It gets corrected, and significant corrections are recorded on the About page rather than quietly edited out. The section there listing what the site previously got wrong is deliberate — a trust page with nothing to admit is less believable than one that admits something.
Are figures like compression ratios and speeds ever estimated?
No. Where a number would be helpful and cannot be verified, the page says so and points at the authoritative source instead. There are no invented percentages, speeds, accuracy rates or platform limits in any tool page on this site.
Do you accept paid guest posts or link placements?
No, and the enquiries do arrive. Nothing on this site is published in exchange for payment, a link or a reciprocal mention. There is no guest post programme and no plans for one.
How do I report something that looks wrong?
Email hello@omniwebkit.com with the page and what you expected. A specific report — the tool, the file type, what happened — is far more actionable than a general one, and reports about inaccurate claims get dealt with before feature requests.
Related pages
About names who runs the site and what it will not do. Privacy covers what happens to files and what the advertising cookies collect. Terms sets out acceptable use and the limits of what the tools promise.