Editorial Policy

How AI is used in producing this publication

AI may help discover sources and prepare first drafts. It does not decide what gets published. A human editor checks and approves every article before it goes live, and nothing generated or imported by AI publishes automatically — every change reaches the live site only through a reviewed, human-approved pull request merged to main.

Because of that, articles here carry a house byline ("The Only Editors") instead of a person's name. It would be misleading to put an individual editor's name on AI-drafted copy as though they wrote it themselves. The "Reviewed by" credit on each article is the real, accountable human who checked it before it went live.

Content types

We publish 4 kinds of articles:

Brief

A concise, sourced update on something that changed.

Test

A hands-on assessment of a tool or feature.

Workflow

A repeatable, step-by-step tutorial.

Pulse

A synthesis of real community discussion and sentiment, not a single sourced change.

How we source a Workflow

A researched Workflow is never a rewording of one person's specific published process, however well credited. It's a synthesis across multiple independent sources covering the same or closely related technique, written in our own words, going beyond any single source by calling out where approaches disagree and adding depth from documented tool knowledge those sources didn't individually cover. When a technique genuinely traces back to one creator's write-up, we credit them and link to their original for the full step-by-step rather than reproducing it here, so our coverage doesn't take the traffic a full rewrite would.

An idea Workflow is different again: it's an automation or combination we've assembled ourselves from tools' real, documented capabilities, not lifted from any existing tutorial because nobody's published that exact combination. We say so plainly near the top of the piece, source every individual capability it depends on, and never claim it's been hands-on verified end to end.

Evidence labels

Every article carries one of three evidence labels, shown on its card and at the top of the article:

Announced

Reported by the maker. Not yet independently checked.

Researched

Checked against sources, but not hands-on tested by us.

Tested

Hands-on tested by our editorial team.

Which levels actually apply depends on format. A Pulse is always Researched, enforced by the schema, since it synthesizes public discussion rather than reporting one sourced change. A Brief is Announced or Researched. A Workflow is, as a matter of current practice, always Researched too, we haven't yet published one we hands-on tested ourselves, though the schema allows it once we do. Tested is reserved for a genuine hands-on Test article.

Effort to try

Every article's sidebar carries an effort rating, except a Pulse, which has nothing to "try" and shows an overall-mood reading instead, see below. It's about time investment, not skill, a tool you already know well might be "Quick" for you even if we rated it "Moderate" for a typical reader coming to it cold:

Quick

Read it and you already have what you need. Nothing to set up or learn.

Moderate

Worth a focused session: some setup, a new interface, or a workflow to get used to.

Involved

A real time investment: new tooling, a multi-step process, or a skill that takes practice.

Pulse: aggregate sentiment, not individual quotes

A Pulse reports what designers, as a group, are actually saying, gathered from spaces like Reddit, Dribbble and Behance comments, and Designer News, deliberately not from engineer-audience forums like Hacker News, which would skew the read toward tech-industry opinion rather than design-practitioner opinion. It never names or quotes anyone individually, even when a claim traces back to one specific comment, the body reads as a synthesis of a pattern, not a transcript of who said what. We only call something a pattern once we've found more than one independent source saying something similar, every one of them linked in the article's sources for a reader to verify, even though none are named in the prose.

In place of an effort rating and the "before you use it" checklist below, a Pulse's sidebar shows two readings of its own. The overall mood is one of Optimistic, Mixed, Skeptical, or Frustrated. How solid the pattern is gets its own reading too:

Emerging pattern

Just starting to show up in a few places, not yet a broad consensus.

Widespread pattern

Showing up repeatedly and independently across multiple spaces.

Before you use it

Every article's sidebar except a Pulse's includes a "Before you use it" checklist covering client data handling, usage rights, disclosure obligations, accessibility, and cost, alongside the tools it discusses and (for Test and Workflow articles) a quick summary. Each item is marked with one of three labels, so you can tell at a glance what we confirmed and what we didn't. We never guess to fill a gap.

Known

Confirmed from the maker's own terms or docs, or a reliable source we cite.

Not verified

We couldn't confirm this. Check it yourself before you rely on it.

Not applicable

This question doesn't apply to what the article covers.

Images, video, and animation

Articles are drafted as text. Any image, GIF, video, or animation is added by a human editor during review, and we only use media we made ourselves, have permission to use, or that its owner offers for embedding (like a YouTube video's own embed). Media from someone else is credited underneath it. Every image and animation has a text description for people using screen readers, and animations pause for readers who've asked their device to reduce motion. Videos don't load anything from YouTube until you press play.

Corrections and withdrawals

When new information changes our assessment, we update the article and bump its "updated" date; if we re-check the underlying facts without changing the conclusion, we bump "last verified" instead. When something we reported turns out to be wrong, no longer applies, or is withdrawn by its maker, we withdraw the article: its URL keeps working, but the page shows a short withdrawal notice explaining why instead of the original body. A withdrawn article is removed from listings, search, and our RSS feed, and marked so search engines don't index it.

Spotted an error, a claim that's out of date, or a source that doesn't say what we said it does? Email theonly at bemosu dot com with the article and what's wrong, and we'll check it against the sources.