QA Intelligence for AI Agents
MCP tools that give AI agents the power to diagnose browser errors, analyze performance, extract structured data, and generate test strategies — all via the Model Context Protocol.
Get notified when the first TestScout MCP tools ship.
Open source on GitHub. Free tier for local use. Paid tier for hosted execution.
Join the Waitlist
Be the first to know when TestScout launches. Get early access and exclusive updates.
Playwright MCP = browser automation (hands). TestScout = QA intelligence (brain).
6 MCP tools shipping sequentially — each encodes real QA expertise that AI agents can use directly.
lightscout-mcp is available now on npm.
Built by QA engineers, for AI agents
QA Superpowers for AI Agents
Six MCP tools that encode real QA expertise — so your AI agents can diagnose, analyze, extract, test, and maintain like a senior QA engineer.
Error Diagnosis
AI agents capture and triage browser errors, console messages, and network failures with full context — no manual debugging needed.
Performance Analysis
Measure Core Web Vitals (LCP, FCP, TTFB, CLS, INP) and get actionable recommendations to improve page speed and user experience.
Structured Extraction
Extract structured data from any web page using natural language instructions. AI agents get clean, typed data instead of raw HTML.
Test Planning & Scaffolding
Describe what you want to test in plain language. TestScout evaluates your app, produces a strategic test plan and a context config — then your AI agent writes the actual test code using any browser MCP.
Load Test Analysis
Generate load tests from plain language, execute them, and analyze results — all within the AI agent workflow via MCP.
Self-Healing Test Suites
Maintain existing test suites automatically. Parallel agents diagnose failures, heal selectors, adapt to changes, and commit fixes — grounded by a context config, not guesswork.
Playwright MCP + TestScout = Complete QA
Playwright MCP (the hands)
- Navigate pages and click elements
- Fill forms and interact with UI
- Take screenshots and snapshots
- Execute browser automation scripts
TestScout MCP (the brain)
- Diagnose errors and triage issues
- Analyze performance and CWVs
- Extract structured data intelligently
- Generate test strategies and load tests
What AI Agents Can Do with TestScout
Give your AI agents the QA expertise they need to debug, analyze, extract, and test — all via the Model Context Protocol.
AI-Assisted Debugging
Let Agents Diagnose Production Errors
Point your AI agent at a failing page. TestScout captures errors, network failures, and console output — the agent triages and suggests fixes.
- Automatic error capture with full stack traces
- Network failure detection and diagnosis
- AI-generated fix suggestions with context
Performance Optimization
Automated CWV Analysis and Recommendations
AI agents measure Core Web Vitals, identify bottlenecks, and generate specific recommendations to improve scores.
- All 6 Core Web Vitals measured automatically
- Actionable optimization recommendations
- Track improvements across deploys
Intelligent Data Extraction
Natural Language Web Scraping
Tell your agent what data you need in plain English. TestScout extracts structured, typed data from any page — no selectors required.
- Describe data needs in natural language
- Get clean JSON output, not raw HTML
- Works on any website without configuration
AI-Driven Test Planning
From Plain Language to Full Test Suites
Describe what to test in plain language. TestScout evaluates your app, produces a strategic test plan and context config, then your AI agent scaffolds the actual test code using any browser MCP.
- Evaluates your app and builds a strategic test plan
- Generates testscout.config.yaml for ongoing maintenance
- Works with Playwright, Vibium, or any browser MCP
Ready to Give Your AI Agents QA Superpowers?
Join the waitlist to get notified when the first TestScout MCP tools ship.
Join the WaitlistAlso check out LightScout
Core Web Vitals analysis, two ways: a Chrome extension for real-time browsing metrics, and an MCP server for AI-powered performance audits via Lighthouse. Both free and open source.
Join the Waitlist
Be the first to know when TestScout launches. Get early access and exclusive updates.
Frequently Asked Questions
Everything you need to know about TestScout MCP tools
MCP is an open protocol that lets AI agents connect to external tools and data sources. TestScout MCP tools give AI agents QA capabilities — error diagnosis, performance analysis, data extraction, and test generation — that they can invoke directly during conversations or autonomous workflows.
They're complementary. Playwright MCP handles browser automation — navigating pages, clicking elements, filling forms. TestScout MCP adds the QA intelligence layer — diagnosing errors, analyzing performance, extracting structured data, and generating test strategies. Think of Playwright as the hands and TestScout as the brain.
TestScout works with any MCP-compatible AI client. This includes Claude (via Claude Code and the Claude desktop app), ChatGPT (with MCP support), Gemini, Cline, Cursor, Windsurf, and any other tool that speaks the Model Context Protocol.
Yes, TestScout MCP tools are open source on GitHub. You can use them locally for free. We'll also offer a paid hosted tier for advanced analysis, execution history, and cloud-based execution — but the core tools are and will remain open source.
The suite includes: testscout-diagnose (error triage and capture), testscout-perf (Core Web Vitals analysis), testscout-scrape (structured data extraction), testscout-plan (test strategy and config generation), testscout-load (load test analysis), and testscout-maintain (self-healing test suite orchestration). They're shipping sequentially, starting with testscout-diagnose.
Without TestScout, AI agents can only reason about errors you paste into them. With TestScout MCP tools, agents can actively capture browser context, measure performance metrics, extract page data, and generate test plans — all programmatically. It's the difference between reading about a bug and investigating it directly.
The free tier includes all core MCP tool functionality running locally on your machine. You get full error diagnosis, performance analysis, data extraction, and test strategy generation. The paid tier adds hosted execution, advanced analysis, trending/history, and team features.
testscout-diagnose is shipping first in 2026, followed by testscout-perf, testscout-scrape, testscout-plan, testscout-load, and testscout-maintain in sequence. Join the waitlist to get notified as each tool launches.
Absolutely. TestScout is open source and we welcome contributions. Once the GitHub repository is live, you can submit issues, feature requests, and pull requests. Join the waitlist to be notified when the repo goes public.
Free users get community support through GitHub issues and discussions. We also provide comprehensive documentation for each MCP tool. For questions or partnerships, reach out at admin@testscout.dev.
Still Have Questions?
Can't find the answer you're looking for? Reach out and we'll help.