Method · ownership · what we will not do

How a screen test is made

What that means for our verdicts

The people who wrote these reviews work for the same studio that made the tools. Nobody should read a site like that as independent, and this one does not claim to be. What we can do is make every verdict checkable and put the weak points where you cannot miss them.

• Every claim about a labs.llc tool points at its source — a file and line number in the build — or at a measurement with the UTC time it was taken.
• Every review has a Dropouts segment naming at least one real shortcoming, and every comparator gets a line on what it does better.
• We did not choose the tools. The ten on this site were drawn at random from the labs.llc register on 2026-09-27 at 19:55 UTC, and the seed was recorded.
• We publish no scores, no stars and no rating markup. The verdict is a paragraph you can argue with.

What we cannot remove is the choice of comparators, and the fact that we know these tools from the inside. Weigh the verdicts accordingly.

Off screen: ė.tv reports its page views to Google Analytics 4 under its own ID, and the tag stores Google’s _ga cookies on your device.

The capture

Each tool was captured from a local copy of labs.llc build 537, byte-for-byte the same as the live build, at 1280 × 800 and 390 × 844 pixels, between 20:13 and 20:16 UTC on 2026-09-27. The time on each frame is the capture file’s own stamp. On the Network review the page’s own clock, visible in the frame, agrees with that stamp to the second. We show captures of labs.llc tools only — never of a competitor’s product.

The callouts

Pins are numbered in reading order and each one is explained in the list beside the frame; the list is the full text version, so nothing is lost if the picture does not load. The “Clean feed” button, where scripts run, hides the pins so the capture can be seen unmarked. A callout describes what is on the screen; where it quotes the page, it quotes it exactly.

The measurement

We read each tool’s source in the build and measured it against the local copy, and read the public feeds it depends on directly, recording the UTC time of every reading. We did not load the live labs.llc server for these tests. Where a test could not be run — a server rewrite our local copy does not apply, a scheduled job that does not run locally, a conversion that needs a real browser — the review says so rather than guessing.

The comparators

Two or three real, current tools per review, checked on 2026-09-27 by fetching their own pages. Where a site refused an automated fetch or returned no readable text, we verified it through search from its own pages, its publisher’s pages, its source repository or its app-store listing, and the review says which. We describe what each does and what it does better; we never guess prices, and we do not claim a competitor cannot do something we did not check. Their links carry no endorsement from us and pass none.

The end card

Each review closes with a verdict in words and a link to try the tool yourself. There is no score, because a number from a reviewer who owns the thing being reviewed would carry more authority than it has earned.

Corrections

If a figure here is wrong, write to labs@labs.llc with the page and the line. We fix it and note the date on the page.