On this page
Editorial standard
photoedit.in publishes an editor, a set of file utilities, and written material that explains how to use them and how the wider tools they compete with actually behave. This page sets out how that written material is produced, what it rests on, and what you can hold us to. It exists because a reader is entitled to know whether an article was researched or assembled, and because we would rather state our standard than have it inferred.
What we publish, and what we do not
We publish three kinds of writing. Task guides explain how to complete a specific job with a specific tool, including the checks a reader should perform on the result. Reference material documents a product's capabilities and limits in depth, of which the twenty-one-chapter Photopea reference is the largest example. Explanatory articles cover the background a task depends on — why a format behaves the way it does, why a file size grows, why a colour conversion loses information.
We do not publish pages whose only purpose is to occupy a search result. We do not generate articles by spinning other articles. We do not publish invented statistics, invented author biographies, invented testimonials or invented customer counts. We do not claim a tool can recover information that was never in the source file, and we do not describe image blur as secure redaction.
Evidence tiers
Every factual claim in our reference material is assigned to one of five tiers, and the tier determines how the claim is written. This is the same scheme published in the sources chapter of the Photopea reference.
| Tier | What it covers | How it is written |
|---|---|---|
| Primary — documentation | The product's own documentation, API reference and release notes | Stated directly, with the documentation as the implicit source |
| Primary — direct testing | Behaviour we observed ourselves against fixed test files | Stated directly, with the test conditions described |
| Primary — vendor statements | Replies from a developer or maintainer in a public issue tracker or forum | Attributed to the statement and dated |
| Secondary — independent sources | Reviews, tutorials and practitioner accounts | Used to corroborate, and described as corroboration |
| Tertiary — reference works | Encyclopaedic entries and aggregators | Used only for dates and background, never for capability claims |
Where a claim is an inference rather than a sourced fact, we say so in the sentence. The self-hosting discussion in the Photopea reference is an example: two sources conflict, neither is decisive, and the article states the most plausible reconciliation and labels it as an inference rather than presenting it as established.
How we test
Testing is done against a fixed set of files rather than whatever happens to be at hand, so that results are comparable between review passes. For image work that set includes a layered PSD with groups, masks, clipping masks, layer styles and editable text; a CMYK TIFF; a camera RAW file; a Sketch document; a Figma export; an INDD package; a 16-bit PNG; and a set of deliberately awkward JPEGs with EXIF rotation and embedded profiles.
Each test records what actually happened, not what the documentation said would happen. Where the two diverge, the observation is recorded and the divergence is noted in the article. Testing is performed on more than one browser engine, because a great deal of browser-based software behaves differently in Chromium than in Firefox or Safari, and that difference is often the most useful thing a reader can be told.
We do not test on a schedule that guarantees freshness. We test when a product ships a change that could affect a claim, and we run a full pass when a reference chapter is due for review. The review date on a page is the date of the last check against the live product.
Review cycle and dating
Every reference page carries three dates in its metadata: when it was first published, when it was last modified, and the human-readable review date shown at the top of the page. The review date changes only when the claims have been checked, not when the page has been edited for style. Where a page has been reviewed by a second person, the byline says so; where it has not, the byline does not imply a second reviewer.
Reference chapters are reviewed individually rather than as a set, because a change to one area of a product rarely affects all of them. That means chapters of the same reference can carry different review dates, and that is intentional rather than an oversight.
Corrections
If something we published is wrong, out of date or unfair to a product, write to contact@photoedit.in with the page and the specific claim. A correction that can be verified is applied, and the review date on the affected page changes. Where a correction changes a conclusion rather than a detail, the change is described on the page rather than made silently, so that a reader who relied on the earlier version can see what moved.
We do not remove a page because it has become inconvenient. If a page is superseded, it is either updated or marked as superseded with a link to its replacement. We do not quietly delete content that has been cited elsewhere.
The most useful reports name the page, quote the sentence, and say what the correct position is with a source we can check. Reports of the form "this is wrong" without a page or a claim are hard to act on. Reports about a product's behaviour are most useful when they include the browser, the file type and the version, because behaviour varies by all three.
Use of AI assistance
We use software tools, including language models, in research and drafting. We do not publish machine output as fact. Every claim that appears in a reference page has been checked by a human editor against a source or against observed behaviour, and the evidence tiers above describe what that check consisted of. Where a tool helped assemble or structure material, a human decided what to keep, what to cut and what to verify.
We do not use AI to generate author identities, reviews, testimonials or statistics, because those things either exist or they do not. We do not publish AI-generated images as if they were photographs of real products or real people. Where an illustration is synthetic, the caption says so.
This disclosure exists because a reader should be able to tell the difference between a page that was researched and a page that was generated. If you believe a specific page falls short of this standard, the correction address above is the right place to raise it.
Advertising and independence
photoedit.in is free to use and is funded by advertising delivered through Google AdSense. Advertising has no influence on what we publish. We do not accept payment for a favourable review, we do not publish sponsored articles presented as editorial, and we do not adjust a verdict because an advertiser might prefer a different one. Where a page compares our own tools with a third-party product, the comparison is written to the same standard as any other, including where the third-party product is better.
We do not run affiliate links in reference material. If that changes, the link will be labelled as an affiliate link at the point of the link, not buried in a disclosure page.
Advertisements are served by Google and are subject to Google's own policies. They do not have access to the images you edit, because images are processed in your browser and are never uploaded to our servers. The privacy policy describes exactly what is and is not collected.
Accessibility
Pages are built to be readable without JavaScript for the content itself, to work at increased text sizes, and to remain usable with a keyboard. Tables carry header cells so that screen readers can associate data with its column. Where we become aware of an accessibility failure on a page, it is treated as a defect rather than a nice-to-have.
How to contact the editorial team
Corrections, accessibility reports, factual disputes and requests for clarification all go to contact@photoedit.in. Technical bug reports about the editor or the utilities should go through the contact page so that they reach the right place. We read everything; we cannot always reply to everything, but a report that names a page and a claim will get a response.
If you are citing our work, cite the specific page rather than the site, because pages are updated independently. Each reference chapter carries a canonical URL and machine-readable article metadata for that purpose.