All You Need to Know About Scout
AI-Powered Autonomous Testing — The Complete Reference
Live & Shipping
scoutqa.ai
1 What Is Scout?
Paste a URL. Describe what to test. Get answers.
Scout is an AI testing platform where you describe testing goals in plain English and AI agents navigate your app like a real user — then deliver structured quality reports.
Scout is an AI testing platform where you describe testing goals in plain English and AI agents navigate your app like a real user — then deliver structured quality reports.
The Core Insight
As AI generates more code faster (Cursor, Replit, Lovable, Bolt), independent validation matters more. The AI that writes your code shouldn’t be the one testing it. Scout is the building inspector for the AI-coded world.
Who It’s For
- ⚡ Indie hackers — launching MVPs and side projects
- ✨ Vibe coders — building with Cursor, Replit, Lovable, v0
- 💻 Small dev teams — no dedicated QA resources
- 🏭 Agencies — client projects on tight timelines
- 📈 Product managers — health check before a demo
Built By
Katalon (100K+ teams globally) — standalone PLG product creating a new category called “vibe testing” — validating not just if software works, but if it feels right.
2 Four Testing Modes
| Mode | What It Does | Best For |
|---|---|---|
| Quick Scan | Fast health check — broken links, spelling, layout, a11y | Pre-launch sanity |
| Deep Run | Multi-pass exploration of user journeys across pages and flows | Thorough testing |
| Intent Mode | Natural language — Scout autonomously figures out how to test | Targeted validation |
| Roast | Brutally honest business + technical critique, shareable | Viral growth loop |
What You Get Back
- ✓ Traffic-light scoring — Green/Yellow/Red per page
- ✓ Issues by severity — Critical, Major, Minor, Info
- ✓ Categories — accessibility, security, performance, usability
- ✓ Screenshots — visual evidence for every finding
- ✓ Fix suggestions — AI-generated remediation steps
- ✓ Shareable links + Markdown export
Intent Mode Examples
3 Use Scout Anywhere (3 Surfaces)
| Surface | How It Works |
|---|---|
| 🌐 Web App | Chat UI at app.scoutqa.ai — paste URL, type intent, watch in real time |
| ⌨ CLI | CI/CD ready — structured output, exit codes, streaming |
| 🔌 MCP Server | One-line config for Claude Code, Cursor, Windsurf, VS Code |
CLI Setup
Claude Code Skill
Cursor Rules
Also Supports
- ‣ Webhooks — notify external systems on completion
- ‣ REST API — /api/v1/ for executions, issues, credits
- ‣ ngrok / Cloudflare Tunnel — test localhost from the cloud
- ‣ Kiro Power — testing integration for Kiro IDE
4 Multi-Agent Architecture
Seven specialized AI agents work together on every test run, built on the Mastra framework.
Orchestrator→Intent Classifier→Planner→Vision AgentNon-Vision Agent→Issue Triage→Report Gen
LLM Configuration
| Agent | Model | Provider |
|---|---|---|
| Orchestrator + Non-Vision | Claude Haiku 4.5 | AWS Bedrock |
| Planner + Vision | Claude Sonnet 4 | AWS Bedrock |
| Fast Classification | Gemini 2.5 Flash |
Tech Stack
Next.js 15React 19TypeScriptMastraAG-UICopilotKitPlaywrightPostgreSQLPrismaAWS EKSStripePostHogLangfuseInngestJotai
Codebase
pnpm monorepo • 10+ packages • 1,040+ PRs • TypeScript strict mode • Node.js 22+
5 What’s New (Recently Shipped)
| Feature | What Changed |
|---|---|
| Intent Workflow | Auto-classifies user intent → routes to optimal testing workflow |
| Credit & Billing | Full Stripe integration: ledger, usage tracking, upgrade modals, webhooks |
| Time-Based Credits | 1 credit = 10 seconds of execution. Complexity-aware pricing |
| MCP Server | Scout as remote MCP server for IDE coding agents |
| Cloud CLI | @scoutqa/cli on npm — CI/CD ready, streaming, structured output |
| App Graph | Knowledge graph accumulates app understanding across runs |
| Recurring Issues | Auto-surface issues that keep appearing across test runs |
| Roast v2 | Business + Technical dual-section format for richer feedback |
| Severity Rubric | Recalibrated triage — fewer false positives, sharper actionability |
6 Pricing & Credits
1 credit = 10 seconds of execution time. Higher models or more complex tasks consume more. Transparent, predictable, fair. Anonymous-first — try Scout without signing up.
Vibe (Free)
$0
300 credits / week
Resets weekly
Ship (Pro)
$15/mo
3,000 credits / month
+ automation & scheduling
Custom
Let’s Talk
Custom credit amount
Dedicated support
All Plans Include
All testing modes, reports, fix suggestions, shareable links, CLI, MCP, webhooks. Plans differ only by credit amount and support level.
Referral Program
Earn credits for inviting friends. Referral credits stack with plan credits (2-month expiry). No rollover on base credits.
7 Roadmap — Near-Term (Q1-Q2 2026)
- ● Vibe Score — per-URL quality tracking over time
- ● Recommended Next Tests — AI-suggested follow-ups
- ● Watch / Monitoring — continuous regression alerts
- ● CI/CD Integrations — GitHub Actions, GitLab CI
- ● Linear Agent — sync issues and test results with Linear
- ● Referral Program — credit rewards for invites
- ● Intent Onboarding — purpose-driven first run
8 Competitive Landscape
| Competitor | Their Approach | Scout’s Edge |
|---|---|---|
| Octomind (a16z) | LLM generates Playwright scripts | True vibe testing — no scripts, pure intent |
| Mabl | Enterprise AI-assisted test automation | 50x cheaper, zero-setup, anyone can test |
| QA.tech | LLM wrapper for Playwright execution | Indie-friendly, instant results, no code |
| Stably | Auto-generates Playwright test code | No test scripts to maintain — ever |
The pattern: Every competitor generates automation code (Playwright) and executes it — they’re LLM wrappers for scripted testing. Scout owns “vibe testing” — true autonomous exploration with zero scripts. We’re opening testing to the citizen tester: everyone can test, not just engineers.
9 Long-Term Vision — Two Products, One Mission
- ‣ Fully autonomous — AI does everything
- ‣ URL + natural language = instant results
- ‣ PLG freemium, $15/mo pro tier
- ‣ Self-serve, viral loops, referrals
- ‣ Indie hackers, solo founders, small teams
- ‣ True autonomous agents that scale with your team
- ‣ Production traffic monitoring for quality insights
- ‣ Ambient QA — 24/7 continuous quality monitoring
- ‣ Progressive learning — app-specific intelligence
- ‣ Enterprise QA teams, large organizations
The Bigger Bet: AI writes most of the code. Scout is the building inspector for the AI-coded world. Quality becomes a utility, not a project — continuous, ambient QA that watches your app 24/7 and catches regressions before users do.
10 By the Numbers
5,293+
Test Executions
1,040+
Pull Requests
7
AI Agents
10+
Packages
3
Surfaces
$15
Pro Plan/mo
Team
7 people (4 Katalon + 3 contractors). Backed by Katalon (100K+ users). Only tool combining intent-based testing + no-code accessibility at $15/mo.