Not a cut-down version
You get the real thing — the SDK in your app, recorded sessions, bug reports with severity, the fix loop. Credits buy tester seats, not a demo tier.
Public beta · free while it lasts
Shipped 90% with AI? The last 10% needs real people.
Crowd testing for AI-built apps — real human testers in 48 hours.
Solo devs here test each other’s apps — because you will never find your own bugs.
You know where everything is, so you cannot see what a stranger sees. Friends install once, tap around, then ghost. Here you test somebody else’s app, earn credits for the testing you actually did, and spend them on testers for your own — no money involved. If you would rather pay than test, that works too.
How it works ↓ Drop-in SDK · Flutter on pub.dev, Expo on npm
Pay only for validated sessions
Privacy-first · GDPR-aligned
Feedback in 24–48 hours
The testing exchange
You cannot test your own app. Somebody else can.
Builders here test each other's apps. You earn credits by testing someone else's app and spend them on testers for your own — the way to clear Google's 12×14 closed-test requirement without paying for it. Same platform, same SDK, same evidence: the only thing that changes is what you pay with.
Earned, not swapped
A test day only pays when a real session backs it, and never on your own app. That is the whole difference to the swap groups: here the testing has to have happened.
Time or money, your call
Got a launch date? Pay and get testers now. Got time instead? Test for two weeks and earn a seat. Both run on the same platform, so switching costs you nothing.
Where this stands today
The exchange opened in August 2026 and is still small — you would be among the first in it, which means finding a match takes a little patience. And a fair warning: Google introduced the 12×14 rule to stop rings of developers waving each other through. Reciprocal testing is fine as long as the testing is real, which is exactly the part we measure and keep evidence of.
The 90-90 rule
The first 90% is nearly free now. Your app is decided in the last 10.
Old engineering rule, truer than ever: the first 90% of the code takes 90% of the time — the remaining 10% takes the other 90% (Tom Cargill, Bell Labs). AI collapsed the first part into days. What is left is exactly the part no prompt can skip: the edge cases, the crashes on someone else's phone, the flows only a stranger gets lost in.
Audiences are tired of demos
Every week another polished prototype launches, trends for a day and goes quiet. A shiny demo no longer proves anything — real usage does.
The last 10% can't be prompted
It is found, not generated: real people on real devices, misreading your UI in ways you cannot imagine. That is what turns a prototype into a product.
Ship proof, not promises
Validated sessions, a bug-to-fix loop and a public evidence profile. "Really used, bugs found, bugs fixed" beats any launch video.
Who's it for?
One platform. Five ways in.
Pick your role. One thought per step — and at the end you decide whether you join.
Step 1 · Your problem
You cannot test your own app.
You walk the paths you built, in the order you built them. The one moment you never reach is the moment a stranger lands on the wrong button.
And friends? They install once, tap around politely and never come back. That is not a test, that is good manners.
Step 2 · Works today
Two lines — then evidence comes back.
flutter pub add testerpaykitnpx expo install @testerpaykit/expo - Every session returns as evidence. Screens, duration, taps, where they gave up — not “works on my machine”.
- Bug reports with proof. Annotated screenshot, voice note, console and network, severity.
- The exchange. You test someone else's app, earn credits, spend them on testers for your own.
Step 3 · Not yet
Three things we do not promise.
- The paid tester pool is empty. Whoever needs testers today gets them through the exchange — or not at all.
- Swift and web are not published yet. Both are built and in use here, but only Flutter and Expo are publicly installable.
- Prices are not fixed. Free during the beta; what applies afterwards we will say before anyone pays.
We write this down because otherwise you find out after signing up — and are rightly annoyed.
Step 4 · Nobody else has this
Approve a bug on your phone. Your Mac builds the fix.
A tester reports a bug. You see it on the go and tap approve. At home a local agent takes the job, works in your repo and leaves you a branch. Your code never leaves your machine.
01
Tester reports
With a screenshot, severity and the path that got them there.
02
You approve
On your phone, out of the queue, in two seconds.
03
Your Mac works
The runner picks up the job and starts the agent locally.
04
A branch to review
Never pushed, never merged. You decide.
Five guard rails are built in — the agent runs with --disallowed-tools Bash(git *), so “never pushes” is a mechanism and not a promise.
Step 5 · Join in
I want in.
Two ways. The first one works today, the second costs you ten seconds.
I will test for someone now
You go into the exchange, take a stranger's app, test it properly — and earn the credits for your own. Give first, then take.
Join the exchangeTell me when it is ready
We write when the paid pool opens or when your SDK lands. No newsletter, just those two messages.
Free while the beta runs. No starting balance — you earn credits by testing for others.
Step 1 · Your problem
You find bugs everywhere. Nobody pays you for it.
You spot the broken flow, the button that lies, the form that eats what you typed. Then you write a review nobody reads, or you delete the app. Either way the work you did is gone.
Testing is a skill. Right now it mostly leaks away into one-star reviews.
Step 2 · Works today
Real apps, your own phone, and your work is on the record.
- You test on the device you actually use. No lab, no scripted click farm. The app runs on your phone, the way a user would run it.
- Every session is recorded as evidence. What you did, how long, where you got stuck. Your work is provable, not a claim.
- Credits for proven work. A test day only counts when a real session backs it — and never on your own app.
Step 3 · Not yet
What you cannot earn here yet.
- Paid jobs do not exist yet. Nobody has bought tester seats so far. Today you earn credits — euros follow when the first project pays.
- You cannot sign yourself up. There is no open registration. You go on the list, and we let you in by hand.
- The selection is thin. The exchange opened in August 2026. Some days there is nothing new to test.
If you are here to earn money this month, say no. This is worth your time if you want to be in early.
Step 4 · Nobody else has this
Your report can become a fix while you are still holding the phone.
You report the bug with proof attached. The developer approves it from their phone, and an agent on their machine starts working on it. You are not shouting into a void — you are the first step of a loop that ends in a branch.
01
You report
Screenshot, voice note, severity — attached automatically.
02
The developer approves
From the queue, on their phone.
03
A fix gets built
On the developer’s own machine, not in a cloud.
04
You see what happened
The bug status says what became of your report.
Step 5 · Join in
Put me on the list.
We let testers in by hand, so this is the only honest button we can give you.
I want to test
You go on the tester list. We write when there is something to test for you — and again when the first paid job exists.
No fee, no obligation, and no minimum hours. If nothing suits you, you test nothing.
Step 1 · Your problem
Your contributors test the happy path. Your users find everything else.
People who work on the code know where the mines are — which is exactly why they cannot walk into them. And there is no QA budget, because there is no budget.
The issue tracker fills with “does not work” and no way to reproduce it.
Step 2 · Works today
Free for open source — a real application, not a coupon.
- Apply with your repo. Email plus a GitHub URL. It goes into a review queue and a human answers.
- Bug reports that can be reproduced. Annotated screenshot, voice note, console and network, plus the path that got there.
- Same SDK as everyone else. Flutter via pub.dev, Expo via npm — publicly installable today.
Step 3 · Not yet
Where this gets thin for a project like yours.
- Free is not automatic. Applications are read by hand. That takes days, not minutes.
- The tester base is small. The exchange opened in August 2026. A niche project may wait a while for a match.
- Only Flutter and Expo are publicly installable. The Swift and web SDKs are built but not published yet.
Step 4 · Nobody else has this
A bug report that ends as a branch in your repo.
Approve the report and an agent on your own machine works it in your checkout, then leaves a branch. Nothing is pushed, nothing is merged, nothing is uploaded — for a project where you are the only maintainer, that is the difference between an issue and a fix.
01
Tester reports
With a screenshot, severity and the path that got them there.
02
You approve
On your phone, out of the queue, in two seconds.
03
Your Mac works
The runner picks up the job and starts the agent locally.
04
A branch to review
Never pushed, never merged. You decide.
Five guard rails are built in — the agent runs with --disallowed-tools Bash(git *), so “never pushes” is a mechanism and not a promise.
Step 5 · Join in
Apply for your project.
Email plus repo. If we say yes, it costs your project nothing.
Free for open source
The application asks for your email and a GitHub URL. A human reads it — expect an answer in days, not minutes.
Apply nowNot ready yet
We write when the tester base grows enough that a niche project finds a match.
Step 1 · Your problem
The client asks whether it is tested. You have screenshots of your own team clicking.
Your people wrote the build. Whatever they find, they find because they went looking for it — and from the outside nobody can tell that apart from a real test.
Step 2 · Works today
Evidence you can forward without editing it.
- Every session comes back as a record. Screens, duration, taps, drop-off reason — a document, not a summary you wrote.
- Bug reports with proof attached. Annotated screenshot, voice note, console and network, severity.
- One SDK across your projects. Flutter via pub.dev, Expo via npm.
Step 3 · Not yet
Honestly: agencies are not served yet.
- There are no agency features. No seats, no white-label, no client-facing report. You would be using the single-developer product.
- The paid tester pool is empty. You cannot buy testers for a client project today.
- Prices are not fixed. There is nothing here you could put into an offer.
If you need this for a client deadline, this is not it yet. Worth ten seconds of your time to know.
Step 4 · Nobody else has this
The fix loop runs on your machine. Client code never leaves it.
Approve a reported bug and an agent works it in your own checkout, then leaves a branch for review. Nothing is pushed, nothing is uploaded to us. For work under an NDA that is not a nice-to-have — it is the precondition.
01
Tester reports
With a screenshot, severity and the path that got them there.
02
You approve
On your phone, out of the queue, in two seconds.
03
Your Mac works
The runner picks up the job and starts the agent locally.
04
A branch to review
Never pushed, never merged. You decide.
Five guard rails are built in — the agent runs with --disallowed-tools Bash(git *), so “never pushes” is a mechanism and not a promise.
Step 5 · Join in
Tell us what an agency would need.
We would rather build the right thing than guess. Go on the list and we will ask you first.
Put me on the list
We write when there is something an agency can actually use — seats, client reports, a price.
Nothing to buy here today. That is the point of saying so.
Step 1 · Your problem
It works in the office. The office is not the world.
Your team tests on company devices, on the company network, with the account that always works. Your users have an older phone, a bad connection and no idea what you meant by that label.
Step 2 · Works today
People who have never seen your app — and a record of every step.
- Real devices, not a farm. The phone someone already owns, with their apps, their storage, their connection.
- Every session comes back as evidence. Screens, duration, taps, where they gave up.
- Bug reports with proof attached. Annotated screenshot, voice note, console and network, severity.
Step 3 · Not yet
What a company would ask for and not get yet.
- No paid tester pool. There is nobody to buy today. The exchange is the only way in.
- No contracts, no SLA, no procurement story. This is a beta run by a small team. Say no if that is a problem.
- Prices are not fixed. Free while the beta runs; what follows we will say before anyone pays.
Step 4 · Nobody else has this
Your source code never leaves your machines.
When a reported bug is approved, the fix is built by an agent on your own developer machine — in your checkout, ending in a local branch. Nothing is pushed, nothing is uploaded to us. Most tools in this category need your repository. This one does not.
01
Tester reports
With a screenshot, severity and the path that got them there.
02
You approve
On your phone, out of the queue, in two seconds.
03
Your Mac works
The runner picks up the job and starts the agent locally.
04
A branch to review
Never pushed, never merged. You decide.
Five guard rails are built in — the agent runs with --disallowed-tools Bash(git *), so “never pushes” is a mechanism and not a promise.
Step 5 · Join in
Tell us what would have to be true.
We are not going to pretend there is an enterprise offer. Go on the list and we will come back with a real one.
Put me on the list
We write when there is a paid pool, a contract and a price — the three things you would need before you could say yes.
Nothing to buy here today. That is the point of saying so.
This is not a mockup. This is what real testing looks like here.
Screenshots from our own local instance, as of June 2026.
Why developers choose TesterPayKit
Everything you need to test your app with real users, without the enterprise price tag.
👥 01
Real testers
A curated pool of verified everyday testers — diverse devices, locations and generations, not the same power-users who already know your app.
📱 02
Real devices
Test on actual phones and tablets, not emulators. Feedback from the devices your users actually own.
⚡ 03
Fast turnaround
Detailed bug reports and structured feedback in 24–48 hours. Ship with confidence.
🔒 04
Privacy-first
GDPR-aligned by default. No PII or raw input content leaves the device unless a tester explicitly attaches it.
✅ 05
Validated sessions
You pay only for sessions that pass: scored for quality, deduplicated, with intentional control-bugs planted to catch click-farmers. No retainer, no minimum.
💰 06
Affordable
Platform-access subscription. You set tester payouts. A transparent transaction fee keeps the platform running.
📊 07
Structured reports
Bug reports with screenshots, widget-level annotations, device info and reproduction steps.
Why we exist
We built TesterPayKit for ourselves first
We're a studio that ships a lot of our own apps — and every one needs real testing before launch. Pro testers are expensive and rare; friends test once and ghost. So we built the tool we needed ourselves. Being customer number one isn't a tagline — it's where the product came from.
Built from real need
Dogfooded on our own apps before we opened it up. If it does not work for us, it does not ship.
Fair to both sides
Developers get affordable real-human testing. Everyday people get a fair way to earn flexible extra income — short tasks, transparent payout.
Honest before launch
No invented numbers, no fake social proof. What is beta, we call beta. Pricing stays placeholder until it is final.
Under the hood
Easy to use. Not easy to build.
From the outside, TesterPayKit looks like a small SDK and a dashboard. Underneath, seven systems have to interlock — built, broken and rebuilt over nine months of daily dogfooding. That depth is the product.
Capture SDKs
Shake-to-report, session capture, breadcrumbs and an offline-safe outbox — native in Flutter, TypeScript/Expo and Swift, all held to a single wire contract.
Evidence pipeline
Every bug arrives with its session: timeline, steps, device context, screenshots, screen recording. Testers do not write essays. Developers do not guess.
Marketplace engine
Job postings, applications, tester validation and reputation — the matching layer between apps that need testing and people who test.
Payout engine
Server-side payout calculation with per-project rules. Real money to real people — including the compliance homework that comes with it.
Two-sided console
One web console, two worlds: developers manage projects, jobs and reports; testers browse, apply and earn.
Agent-native tooling
A CLI and an MCP server feed bugs and sessions straight to AI coding agents — your agent reads the report and ships the fix in the same loop.
API backbone
110+ endpoints carrying auth, roles, magic links, notifications and payouts — the part nobody notices until it breaks.
By the numbers
- 9 months of building
- 110+ API endpoints
- 3 SDKs, one contract
- 1,000+ automated tests
Could someone copy it? Sure — after they rebuild all seven. We will keep shipping in the meantime.
How TesterPayKit compares
We build for indie developers shipping with AI tools. Compared against the two crowd-testing alternatives we get asked about most.
| TesterPayKit | UserTesting | TestFi | |
|---|---|---|---|
| Starting price | €0/month · pay-per-tester | $49–$150 / session | $1.99–$3.99 / tester |
| SDK integration | Optional (for deep telemetry) | None | None |
| AI-built app focus | Yes — Cursor, Claude Code, Lovable, Flutter | Generic | Generic |
| Auto-captured repro steps | Yes (breadcrumbs + crash context) | No | No |
| Paid testers from pool | Yes — DACH focus | Yes (US-centric) | Yes (12+ countries) |
| Hosting | EU (Hetzner) · GDPR-aligned | US | Unknown |
| Agentic fix loop (Claude Code) | Yes — tpk CLI + plugin | No | No |
Sources: vendor pricing pages (as of 2026-05). Feature claims from each vendor’s own marketing site at time of comparison.
Two ways to test. Both honest.
Bring your own testers and pay only for the toolkit — or use our paid pool and pay per run. No seat quotas, no hidden payout cuts.
No prices yet. We are not quoting numbers yet, because we have not set them yet — and a made-up price is still a made-up number. What is settled is the structure: bring your own testers on a subscription, or have us bring paid ones per run. If you want to know what a run would actually cost, the calculator on our pricing page does the maths from your own inputs, and shows what the tester gets as well as what we keep.
Bring your own testers
Toolkit
You already have people — staff, community, a Discord. Pay only for the platform that tracks, proves, and routes the testing. Nobody is paid through us, so there is no payout cut.
Free / OSS
Free
Bring your own testers — staff, community, friends.
- Your own testers, unlimited
- 1 project
- Full console + bug reports
- Session evidence (screens, time)
- 7-day retention
- Free forever for OSS maintainers
Where we would start
Team
For companies testing with their own people.
- Everything in Free
- Up to 10 projects
- SDK + automatic bug capture
- Fix loop into Claude Code / Cursor (MCP)
- Screen recordings + structured reports
- 1-year retention · email support
Business
For agencies and larger in-house QA.
- Everything in Team
- Unlimited projects + seats
- SSO (SAML/OIDC)
- API access & webhooks
- White-label / exportable reports
- Dedicated support · invoice billing
We bring the testers
Managed pool
No testers of your own? We recruit and pay real everyday testers. You pay per test run — their time plus our margin, never less than we pay them.
Gate
Clear the Google Play 12-testers-for-14-days gate.
- 12+ real, paid testers
- Opted in for the full 14 days
- Daily opted-in count
- Pool replacements if someone drops
- Status tracking in your console
Where we would start
Evidence
Know testing really happened — not hope it did.
- Everything in Gate
- Validated sessions: screens, time, what was exercised
- Screenshots
- Short human-readable test report
Bug Hunt
Ship with a fix list, not a shrug.
- Everything in Evidence
- Structured bug reports + repro steps + device data
- UX-friction notes + launch-readiness checklist
- Pull bugs into Claude Code / Cursor via SDK + MCP
🚀
50 % startup discount
First year, for companies under 2 years and < €1M funding.
💚
30 % non-profit discount
For registered non-profit organizations.
🧪
Free during the beta
No one pays anything before final pricing is published.
Frequently asked questions
How much does crowd testing cost compared to UserTesting? +
TesterPayKit starts at €0/month — you only pay testers who actually run your app. UserTesting charges $49–$150 per session. For indie apps with a few test cycles per month, TesterPayKit is multiple times cheaper.
Do I need to install an SDK to use TesterPayKit? +
No. The SDK is optional and gives deeper telemetry (breadcrumbs, crash reports, screen recordings), but you can run tester campaigns and collect feedback without it.
How do testers get paid? +
Testers receive transparent per-task payouts, released after your approval. Rates run through the TesterPayKit platform; testers see what they will earn before they start a task.
Can I crowd-test a Flutter app without my own user base? +
Yes — that’s the whole point. TesterPayKit brings real people (students, retail workers, retirees) as testers for apps that do not yet have an audience.
What is crowd testing for indie developers? +
Crowd testing is a model where a distributed pool of real users tests an app on real devices, instead of in-house QA only. For indie developers, that means realistic feedback without owning a user base.
How long does a test campaign take? +
Typical campaigns deliver first feedback within 24–48 hours. More complex multi-session tests run for 5–7 days.
Is TesterPayKit GDPR-aligned? +
Yes. TesterPayKit is EU-hosted (Hetzner), and we avoid storing tester clear-text data or app content server-side beyond what testing requires. Privacy + DPA are part of the standard setup.
Build with AI. Ship with humans.
The first 90 % is AI. The last 10 % — the bugs real thumbs find — is other builders, testing your app while you test theirs. Free while the beta lasts.
No credit card required · Free plan available · Cancel anytime