Send a task that needs a human. Get the answer in minutes.
SeenByHuman is the human layer for AI workflows: your agent — or you — sends a micro-task (test an app, judge a screen, gather 100 opinions, check if the copy sounds right in a market) and real people respond in minutes. Set a budget; we allocate the right number of humans — $100 → 20×$5, 10×$10, or 4 verified × $25. See the API →
Pick a human from the country & language you need
Not a bot — a real person in that country who runs your app with their own eyes, records the screen and tells you what stopped them. Price known up front. Ratings after every job keep quality honest.
Small jobs for a dollar (or less)
Not everything needs a full test. Sometimes a fresh human eye for a minute is enough — and a tester earns a little without getting off the couch.
Which screen looks finished, and which looks "made by a bot"?
$0.503 questions on the first impression of your app. 2–3 minutes.
$1A native opens the app for 60 seconds and says what stands out.
$1Someone from that market checks the copy doesn't read like a translation.
$1Drop a link, set a budget — a human does the rest
What we test
Matched tester
- Screen recording with voice notes
- Issue list by severity + screenshots
- UX opinion in your market's language + 1–10 verdict
From a link to a report written by a human
You drop the app & budget
A link, two lines on what to check, a budget slider. Price up front — no quotes, no waiting for an offer.
A human runs the app
A matched tester at the level you set goes through your app like a real user, records the screen and says out loud what stopped them.
You get a verdict & rate it
Screenshots, issues by severity, a UX opinion and a 1–10 verdict. You rate the tester — that drives their level and your trust in the next one.
A report a human reads — and your AI agent acts on
"Signup went fine, but I got stuck adding a task — I tapped Save four times, nothing. You can hear me getting lost on the recording. Otherwise it looks really good, I'd show it to a friend."
↳ pinned to 0:41 of the recordingThe full report has a screen recording with voice notes and a screenshot on every issue. You read it like a human; your AI agent gets the same as structured data and fixes the code itself.
Your agent orders the test itself
This is the main way to use SeenByHuman: after it builds the app, your AI agent calls it like any tool — passes a link and a budget, a human runs it, and the verdict comes back as clean data the agent acts on. You click nothing.
tool seenbyhuman.order_test(
url="https://moja-apka.lovable.app",
check="signup, payment, does it feel done?",
budget=20, country="PL", level="verified"
)
# → structured verdict comes back
{ "verdict":7, "issues":[
{ "severity":"critical", "what":"Save button dead on iOS" }
] }
Earn from your couch. No interviews, no CV, no gatekeeping.
Got a phone and a free evening? Sign up in a minute, switch yourself to "available" and grab your first job — a full test or a dollar micro-task. Better ratings mean a higher level and better-paid tests.