What owning both sides means
A guide that recommends its publisher’s own tools has an obvious pull towards saying yes. We push back against it in four ways, and you can check each one on any review.
Good to know before you read: eriks.best measures its visits with Google Analytics 4, in its own property, and Google’s _ga cookies are how that count is kept.
- No ratings, stars or scores anywhere, and no review or rating markup for search engines. A tool is either the right fit for a job or it is not.
- Every claim about a labs.llc tool is traced to its source code in build 537 or to a measurement with its UTC time, and the review lists those sources at the end.
- Every review names at least one real shortcoming — something the tool does not do, or does worse than a rival.
- The “which one to pick” table gives rows to the alternatives whenever they do the job better, and the home page repeats that under “look elsewhere if” on every card.
How we test a labs.llc tool
We read the tool’s own files from build 537 of labs.llc and describe what the code does, with file names and line numbers. Then we measure: each page, script and relay was requested from a copy of build 537 served on our own machine — byte-for-byte the build that is live — and the time of every reading was recorded in UTC. Where a tool calls a public source (a DNS resolver, the USGS, Open-Meteo), we called that source the same way the page does and noted what came back.
Where a claim depends on the tool’s logic rather than its screen — the fifty Mind Sphere versions, the Glass byte walker, the Watermark stripping pattern — we ran the page’s own code in macOS’s JavaScriptCore — on test photographs and a crafted string where it needed an input — so the result is the code’s, not our impression.
How we check the alternatives
Each alternative is a real, current tool on the open web, and we name two or three per review. We fetched each one’s own pages on 27 September 2026 and describe only features we saw there. When a site refused our request or drew its page entirely with script — the Smithsonian’s Global Volcanism Program, Every Noise at Once, Windy.com and the Radiooooo web app — we used its Wikipedia article, its app-store listing or a web search that returned its own pages instead, and the review says so.
We do not guess prices, we do not say a rival “cannot” do something we did not check, and we use no rival logos or screenshots. Links to alternatives are ordinary links marked nofollow.
How to read the cards
- Best for — the jobs and people for whom the tool is our pick, each backed by something in the review.
- Not for — the jobs where it is the wrong choice. Most of these point to a named alternative in the table above them.
- Pick this if… — one sentence for readers in a hurry. If it does not describe you, read the Not for card.
Screenshots
Pictures on this site are screenshots of labs.llc tools only, captured from build 537. When a capture caught a page mid-load — the Mind Sphere canvas, the New York City hero picture — the alt text says so rather than describing what should have been there.
Corrections
If something here is wrong, write to labs@labs.llc and we will fix the review and note the date of the change.