What if you just said

check if the purchase
event actually fires

It completes the purchase flow. Reads the analytics logs. Checks the event name, the payload, the price. Passes or fails.

make the API fail and
see what the app does

It mocks the response. Triggers the error. Checks if the error screen shows. Reads the logs. Tells you what broke.

test my last 3 commits
before I push to main

It reads the diff. Finds changed screens. Generates tests. Runs them. Catches a bug. Files it in Linear with screenshots.

book a ride and check
if the fare is correct

It opens the app. Picks a route. Taps through every screen. Reads the total. Checks the API. Done in 60 seconds.

SCROLL
All of this. One prompt. Free to start.

No test scripts. No selectors. No YAML. Just describe what to test.

Free during beta — unlimited test runs — iOS & Android

Test your mobile app.
In plain English.

Describe what to test. The AI agent navigates your app, finds bugs, and tells you exactly where in your code to fix them. Everything runs 100% locally on your Mac — only the AI (Claude) is cloud.

Download Free Beta See How It Works
NoobQA
+ Library ... Sonnet 4.6 Simulator
Claude
Xcode
Simulator
Noober
iOS Simulator • Android Emulator • macOS Native App • Bring Your Own LLM • 100% Local • No SDK Required (Android) • Real Devices COMING SOON
NoobQA

Get early access

Be the first to try NoobQA. Early adopters get 50% off forever.

No spam. Just a launch notification + your discount code.

How it works

Three steps. No scripts. No setup headaches.

1

Describe your test

Write what you want to test in plain English. "Open checkout, verify the fare, check if the analytics event fires."

2

AI runs the test

The agent opens your app, taps buttons, navigates screens, handles popups, and verifies everything — autonomously.

3

Get a report

Pass or fail with evidence. API verification, screenshots, bug details. Auto-file in Linear, send to Slack.

Scroll to watch the test run

NoobQA
00:00
Home Screen
iPhone 16 Pro
Book a Ride
My Trips
Profile
Setting up...
Simulator
Claude
Noober
148 turns

Get a real report

Pass or fail with evidence. Every test run produces a detailed debrief with bugs, screenshots, and actionable fixes.

NoobQA
Debrief Smoke Test
All Tests Passed
4 Passed 1 Failed 2 Bugs 04:12
TC01
Login
✅
TC02
Browse
✅
TC03
Book
✅
TC04
Payment
❌
TC05
Receipt
✅
Feed
Bugs 2
Report
Payment gateway timeout on slow networks P1 CRITICAL
POST /api/v1/payments returned 504 after 30s. App shows spinner indefinitely with no timeout handling. User stuck on payment screen.
PaymentViewController.swift:142 TC04
Analytics event fires with stale price after coupon toggle P2 MEDIUM
purchase_completed event sends price: 149 even after coupon removed (expected: 249). Event payload not recalculated on toggle.
BookingAnalytics.swift:87 TC04
Claude
Xcode
Simulator
Noober
312 turns

Not just another test runner

NoobQA doesn't replay scripts. It thinks, navigates, recovers from errors, and validates your app like a real human QA — powered by AI.

Autonomous Navigation

Say "book a ride and verify the fare." NoobQA figures out the taps, handles popups, dismisses alerts, and navigates to the target screen.

Noober Deep Inspection

See every API call, response body, WebSocket frame, and app log. Validate API correctness — not just UI appearance.

200x Faster Screen Reading

Most AI tools send screenshots and guess what's on screen. NoobQA reads the actual UI structure in 100ms — not 30 seconds. Faster tests, more accurate results.

API & Analytics Verification

Verify the booking API returned 201. Check if the analytics event fired with the right payload. Catch backend bugs, not just UI glitches.

Record & Replay

Record a user flow once. Every tap, swipe, and screen transition is captured. NoobQA replays it 10x faster — and self-heals when the UI changes.

Self-Learning Agent

The agent learns from every test run. "Toggle needs touch() not tap()" gets saved and injected into future tests. It never makes the same mistake twice.

What a real test looks like

You type one sentence. The AI does everything else.

You type
Test the login flow, book a ride to the airport, and verify the payment goes through
Sonnet 4.6 Simulator
NoobQA does
Launched app on iPhone 16 Pro 0:02
Tapped "Sign In" → entered credentials 0:08
Navigated to booking → selected airport 0:18
Confirmed ride → tapped "Pay Now" 0:25
Verified POST /api/payments → 200 0:30
Bug: analytics event has stale price after coupon toggle 0:33
PARTIAL PASS 2/3 passed · 1 bug filed · 33 seconds

Why not just use XCUITest?

Or Appium. Or Detox. Or any script-based framework.

CAPABILITY XCUITEST / APPIUM NOOBQA
Write a test Code + selectors + setup Plain English
UI changes break tests? Yes — fragile selectors Self-heals — reads screen fresh
Validate API responses No — UI only Yes — intercepts all traffic
Check analytics events No Yes — reads app logs
Handle popups & alerts Manual handling code Autonomous — dismisses them
Debug a failure Read logs manually Files bug with repro + evidence
Setup time Hours (framework + config) 5 minutes
Extensible via MCP

It doesn't just find bugs.
It takes action.

NoobQA connects to your team's tools. File bugs, notify your team, and extend with any MCP server.

Linear

Bug tracking

You say:
"Test the checkout flow and file any bugs in Linear"
NoobQA automatically
✓ Creates issue: "Analytics event fires with stale data after coupon toggle"
✓ Adds repro steps, screenshot, API evidence
✓ Sets priority, assigns to the right team

Slack

Team notifications

You say:
"Run the full smoke test and post results to #qa-reports"
NQ
NoobQA Bot 2:14 PM
✅ Smoke Test — PASS (5/5)
Login ✓ · Home ✓ · Booking ✓ · Payment ✓ · Profile ✓
Duration: 3m 12s · 0 bugs · Build #247

Custom MCPs

Infinite extensibility

Add any MCP server. Your agent gains new superpowers:

⚡
Sentry MCP
"Check if this screen has any crash reports in Sentry"
📊
Mixpanel MCP
"Verify event 'purchase_completed' shows in Mixpanel"
🚀
Your Internal API
"Reset test user data before running the flow"
Settings → Integrations → Add Custom MCP

The full loop — automated

Write test
Plain English
AI tests app
Taps, verifies APIs
Finds bug
With evidence
Files in Linear
With repro steps
Posts to Slack
#qa-reports

Trigger from anywhere

NoobQA fits into your workflow, not the other way around.

From Terminal

URL scheme launches NoobQA from any script or CLI.

$ open "noobqa://run?prompt=verify+checkout&autorun=1"

From Git Diff

Auto-generate tests from your uncommitted changes. NoobQA reads the diff, identifies affected screens, and tests them.

$ open "noobqa://diff?path=./&autorun=1"

From Claude Code

After writing code, Claude Code can trigger a NoobQA test to verify nothing broke — right from your editor.

/noobqa "test the screen I just changed"

Free during beta.

Everything runs locally on your Mac. Only the AI (your Claude subscription) is cloud. No data leaves your machine.

BETA — FREE Unlimited test runs · iOS & Android · No credit card
Bring your own LLM
| Currently: Claude Code | Coming: OpenAI, Gemini, local LLMs

NoobQA uses your Claude Code subscription ($20/mo) or API key as the AI engine. You keep full control of your LLM costs (~$0.01–0.05 per test).

BETA — FREE
Everything Included
$0
free during beta — unlimited
  • ✓ Unlimited test runs
  • ✓ iOS Simulator + Android Emulator
  • ✓ Deep app inspection (network, logs, storage)
  • ✓ Root cause analysis (exact file:line)
  • ✓ AI learns from every run
  • ✓ Mock API responses
  • ✓ Figma visual comparison
  • ✓ Linear + Slack integrations
  • ✓ Android: zero setup (no SDK needed)
Download Free Beta
WHAT YOU NEED
Your setup
100% Local
everything runs on your Mac
  • ●
    Your Mac — NoobQA runs as a native macOS app
  • ●
    Claude Code — $20/mo subscription or your API key. This is the only cloud part — the AI brain.
  • ●
    iOS — Xcode + Simulator (you already have this)
  • ●
    Android — Android Studio + Emulator (no SDK integration needed)

Your code, screenshots, and app data never leave your machine. NoobQA only sends test prompts to Claude.

COMING SOON
Cloud
TBD
monthly · for teams
  • ✓ Everything in Pro
  • ✓ Cloud test runners
  • ✓ Shared test libraries
  • ✓ Team dashboards
  • ✓ Test history + trends
  • ✓ Multiple LLM support
  • ✓ Android testing
Get Notified

What makes NoobQA different?

AI coding assistants can do basic taps on a simulator. Here's what NoobQA builds on top:

Deep App Inspection

See every API request/response, app logs, analytics events, and app state via Noober SDK. No other tool can look inside the app.

Learns & Gets Faster

Saves successful paths for instant replay. Extracts lessons from failures. Reads screens 200x faster than screenshot-based tools.

Full QA Workflow

Structured reports with pass/fail. Auto-file bugs in Linear. Compare against Figma designs. Send to Slack. CI/CD ready.

Competitors charge $120–500/mo for less. NoobQA: free tier + $29 one-time Pro.

Is NoobQA for you?

✓

Perfect for

  • ✓
    Indie iOS & Android developers — no QA team, limited time for manual testing
  • ✓
    Small studios (2-15 devs) — QA bottleneck slowing releases
  • ✓
    Smoke testing before every release — catch regressions early
  • ✓
    API + UI verification together — not just screenshots, sees inside the app
  • ✓
    Teams that hate writing test scripts — XCUITest, Espresso, or Appium → plain English instead
✕

Not the best fit for

  • ✕
    Hardware-dependent testing — camera, Bluetooth, NFC, GPS accuracy REAL DEVICES COMING SOON
  • ✕
    Enterprise device farms — use BrowserStack or Sauce Labs for 100+ device matrix
  • ✕
    Performance / load testing — we test functionality, not throughput

How we compare

NoobQA Maestro Appium Revyl
Plain English tests ✓ ✕ YAML ✕ Code ✓
Network/API inspection ✓ via Noober ✕ ✕ ✕
Learns from past runs ✓ ✕ ✕ ✕
100% local / private ✓ ✓ (CLI) ✓ ✕ Cloud
Real device support ✕ Simulator ✓ ✓ ✓
Price Free / $29* $250/device/mo Free (+ engineering time) Enterprise (est. $500+)

FAQ

Why Simulator and not real devices? +
80-90% of QA doesn't need hardware. API bugs, navigation regressions, UI layout issues, state management bugs — all run perfectly in the Simulator. The things that need real devices (camera, Bluetooth, NFC) are edge cases tested manually anyway. Physical device support is on our roadmap.
What's the Noober SDK? Do I need it? +
No, it's optional. Without Noober: NoobQA tests via screenshots and accessibility tree (like all competitors). With Noober: it sees network requests, app logs, analytics events, and app state — things no other tool can access. Add one line of code to unlock deep inspection. Or don't — basic testing works without it.
How much does Claude API cost per test? +
~$0.01–0.05 per test run with Sonnet 4.6. A typical test uses 5K-20K tokens. NoobQA's subscription covers the tool — you bring your own Claude subscription or API key. This keeps our margins honest and your costs transparent.
What makes this different from Maestro, Appium, or XCUITest? +
Those tools require you to write and maintain scripts. NoobQA is autonomous — you describe what to test, the AI figures out how. It handles popups, login screens, and unexpected states on its own. And it learns from past runs so it gets faster and smarter over time.
Is my code/data sent to the cloud? +
NoobQA itself doesn't collect or store any of your data. The app runs entirely on your Mac. However, since the AI agent is powered by Claude, your test prompts and screen context are sent to Anthropic's API for processing — the same way it works when you use Claude Code in your terminal. We don't see, store, or have access to any of that data. Your code, screenshots, and test results stay on your machine.
Can I use this in CI/CD? +
Yes. We include a GitHub Actions template that runs tests on every PR. It boots a macOS runner, launches the Simulator, runs your test plan, and posts the report as a PR comment. Works with any CI that supports macOS runners.
What's on the roadmap? +
See the roadmap below. We're building in public — follow along on Twitter for weekly updates.

Roadmap

What we're building next. Follow along as we ship.

Now — v1.0

Core Testing Engine

Claude Code agent Noober deep inspection iOS Simulator Android Emulator Zero-setup Android (Frida) Learned lessons Linear + Figma
Next — v1.x

Multi-Agent & Devices

Direct Claude API (no CLI needed) OpenAI / Gemini support Physical device testing Test scheduling (cron) Slack reports
Future — v2.0

Platform Expansion

Team dashboards Local LLM support (Ollama) Shared test libraries Cloud test runners
Get Early Access

Up and running in 5 minutes

Join the waitlist for early access. We'll notify you when NoobQA is ready to download.

NoobQA App

Native macOS · Free trial included

✓ AI-powered autonomous testing
✓ Integrations — Linear, Slack, Figma
✓ Simulator control — tap, swipe, type
✓ Deep API & analytics verification
✓ Test library, learned lessons, flow replay
✓ CLI tool, Git hooks, CI/CD ready
Free tier + Pro $29 one-time (launch) · Bring your own LLM

Noober SDK

Open Source · Optional but powerful

✓ See every API request + response body
✓ Read app logs, analytics events
✓ Switch environments (staging/prod)
✓ Record user flows for instant replay
✓ Semantic screen HTML (100ms vs 30s)
✓ Works without it — but 10x better with it
One line of code: Noober.shared.start()

Get started in 5 minutes

Step 1 — Choose your platform
iOS
Android
// iOS: Add Noober SDK (one line)
import Noober
Noober.shared.start()
// Android: zero setup — Frida auto-injects
// Just select your app in NoobQA. That's it.
Step 2 — Clone & run NoobQA
$ git clone https://github.com/noob-programmer1/NoobQA
$ cd NoobQA && swift run
// or download NoobQA.app from Releases
Step 3 — Test in plain English
▶ Run Test
"Open the checkout, verify the fare breakdown,
and check the analytics event fires correctly"

Works with any iOS or Android app. SwiftUI, UIKit, Jetpack Compose, XML Views. No test framework needed.

Join Waitlist View Pricing

Stop writing tests.
Start describing them.

Early adopters get 50% off forever. One sentence to test.

Join Waitlist