qmax-code is an AI testing agent that lives in your terminal. It ships 50 QA tools — crawl, generate, execute, and heal Playwright and k6 tests — accessible through a natural-language interface. One command to install, zero dependencies, works on macOS, Linux, and Windows.

How is qmax-code different from Claude Code, Cursor, or Windsurf?

Claude Code, Cursor, and Windsurf are general-purpose coding agents. qmax-code is a specialised QA agent: it knows how to crawl web apps, generate Playwright and k6 scripts with real selectors, execute tests in the cloud via the QualityMax Sandbox, and self-heal tests when the UI changes. It runs alongside those tools in the same terminal, not in place of them.

Does qmax-code work with Playwright?

Yes. qmax-code generates Playwright TypeScript tests from natural-language descriptions or from crawling a live URL. Generated scripts use observed selectors (data-test attributes, role queries, text matches) rather than invented ones, and include a self-healing layer that updates selectors when the DOM changes.

Can qmax-code generate k6 load tests?

Yes. qmax-code generates k6 JavaScript tests for load, stress, spike, soak, smoke, breakpoint, API, and browser scenarios. It can also convert existing JMeter, Gatling, Locust, and Playwright scripts to idiomatic k6.

How are API keys stored?

API keys are stored in the OS keychain — macOS Keychain on Apple platforms, libsecret / secret-tool on Linux, Credential Manager on Windows. Keys are never written to disk in plaintext. Masked input with dots is shown while typing.

Does qmax-code cost anything?

qmax-code is included in every QualityMax plan, including the free BYOLLM tier. On the free tier you bring your own Anthropic API key and pay only the LLM usage. Paid plans bundle managed AI credits and add cloud execution.

Is qmax-code open source?

qmax-code binaries are distributed at github.com/Quality-Max/qmax-code-releases. Source for the tool manifest and local integrations is open; the core orchestration is proprietary.

qmax-code v1.5.0 — Terminal

$ qmax-code ██████╗ ███╗ ███╗ █████╗ ██╗ ██╗ ██╔═══██╗████╗ ████║██╔══██╗╚██╗██╔╝ /\_/\ ██║ ██║██╔████╔██║███████║ ╚███╔╝ ( ^.^ ) ██║▄▄ ██║██║╚██╔╝██║██╔══██║ ██╔██╗ > ~ < ╚██████╔╝██║ ╚═╝ ██║██║ ██║██╔╝ ██╗ /| |\ ╚══▀▀═╝ ╚═╝ ╚═╝╚═╝ ╚═╝╚═╝ ╚═╝ meow. code 1.5.0 Mode: standalone (direct API) Logged in as: you@company.com qmax > crawl staging.myapp.com and generate tests█

The first AI testing agent
built for your terminal

Name: qmax-code
Author: QualityMax

No other tool does this. 50 QA tools, natural language interface, self-healing scripts, cloud execution, k6 load testing — all from the command line. Meet Max the cat.

curl -sL https://qualitymax.io/static/install-qmax-code.txt | bash

macOS (Apple Silicon & Intel) · Linux (x64 & ARM) · Windows — 15MB, zero dependencies

Demo

See It in Action

Watch QualityMax generate, execute, and self-heal tests — end to end.

Capabilities

50 tools. One agent. Zero context-switching.

Tell Max what you need in plain English. He orchestrates the right tools, shows progress, and reports results — all without leaving your terminal.

AI Crawl & Test Generation

Point at a URL. Max crawls it with AI vision, discovers pages and interactions, generates Playwright tests with real selectors.

start_crawl, crawl_results

k6 Performance Testing

Create, run, and analyze load tests. Generate from URLs, convert from JMeter/Gatling/Locust. Full k6 lifecycle.

k6_generate, k6_run_test, k6_report

Cloud Test Execution

Run Playwright tests in the QualityMax Sandbox with live progress bar and ASCII browser animation. Screenshots, video, traces.

run_test, check_test_status

Self-Healing Scripts

When a test fails, Max analyzes the error, fixes selectors, and re-runs. Rollback if the heal makes things worse.

update_script, rollback_script

Test Case Management

Full CRUD for test cases and projects. AI-powered enhancement adds edge cases and assertions. Import from docs or TestRail.

create_test_case, enhance_test_case

CI/CD Framework

Export a complete test framework. Get the one-liner install command. Trigger CI runs from the terminal.

export_framework, trigger_framework_run

Screenshots & Vision

/screenshot captures your screen. Drag images into terminal. /paste from clipboard. AI analyzes them with vision.

/screenshot, /paste

OS Keychain Security

API keys stored in macOS Keychain or Linux secret-tool. Never plaintext. Masked input with dots while typing.

/keys, /connect

Syntax Highlighting

Code blocks re-rendered with glamour after streaming. JavaScript, Python, YAML all highlighted in the terminal.

glamour re-render

Workflow

Install. Connect. Test. Ship.

From zero to running tests in under a minute.

Install

One curl command. 15MB binary. No dependencies.

→

Connect

Type /connect. Browser opens. Click authorize. Done.

→

Test

"Crawl my app and generate tests." Max handles the rest.

→

Ship

Export framework. Push to GitHub. Tests run on every PR.

In action

Real sessions. Real results.

Every terminal block below is a real workflow that qmax-code can do right now.

Onboard a new project in 60 seconds

qmax > import https://github.com/acme/checkout-app with base url https://staging.acme.io ✓ Project created: checkout-app (ID: 142) ✓ Repository imported: acme/checkout-app (TypeScript, Next.js) ✓ 23 test cases generated from code analysis ✓ 8 Playwright scripts auto-generated qmax > crawl the staging site and find what we missed ⣹ Crawling https://staging.acme.io... ✓ 14 pages discovered, 6 forms detected, 3 auth flows ✓ 11 additional test cases generated ✓ Total coverage: 34 test cases, 19 automated qmax > export the framework and give me the install command ✓ Framework exported: checkout-app-tests.zip (19 specs) Install in CI: npx @qualitymax/cli install --project 142 --token $QMAX_TOKEN

Run tests with live progress, auto-heal failures

qmax > run all tests for project checkout-app 🌐 ┌──────────────────────────────────┐ │ ◉ ◉ ◉ ▓▓▓▓▓░░░░░░░░░░░░░░░ │ │ ▓▓▓▓▓▓▓▓▓▓▓▓░░▓▓▓▓▓▓▓▓▓▓▓▓▓▓ │ └──────────────────────────────────┘ ⣹ █████████████████████████░░░░░ [32s] 83% Running test 6/8... ✓ 6 passed ✗ 2 failed Execution: exec_1204_1774709000 Status: ✗ FAILED Errors: page.getByRole('textbox', {name: 'Password'}) — not found Screenshots: 2 captured Video: recorded qmax > heal the failures 🔧 Script #1204: Fixed password selector Before: getByRole('textbox', { name: 'Password' }) After: locator('input[type="password"]') 🔧 Script #1207: Updated nav link Before: getByRole('link', { name: 'Sign In' }) After: getByRole('link', { name: 'Log in' }) ✓ Re-run: 8/8 passed [41s]

Load test your API in one sentence

qmax > generate a k6 load test for https://api.acme.io hitting /auth, /products, /checkout ✓ k6 script created: api-load-test (ID: 847) Type: load | VUs: 50 | Duration: 5m | Thresholds: p95 < 500ms qmax > run it ⣹ ███████████████░░░░░░░░░░░░░░░ [2m18s] 50% Ramping VUs... ✓ Load test complete [5m02s] Results: ┌────────────────┬─────────┬─────────┬─────────┐ │ Endpoint │ Avg │ p95 │ Status │ ├────────────────┼─────────┼─────────┼─────────┤ │ POST /auth │ 142ms │ 289ms │ ✓ pass │ │ GET /products │ 67ms │ 134ms │ ✓ pass │ │ POST /checkout │ 384ms │ 712ms │ ✗ fail │ └────────────────┴─────────┴─────────┴─────────┘ ⚠ /checkout p95 (712ms) exceeds threshold (500ms)

Find what you're not testing

qmax > do a gap analysis on the checkout-app repo ⣹ Analyzing repository... scanning 147 files, 12,400 LOC ✓ Analysis complete Coverage Gaps Found: ● Critical — Payment webhook handler (0% coverage) src/api/webhooks/stripe.ts — handles refunds, disputes, subscription changes ● Critical — Cart race condition on concurrent updates src/services/cart.ts:L142 — no locking, shared state mutation ● High — Password reset flow untested src/pages/reset-password.tsx — 3 states, token validation, expiry ● High — Admin panel has no E2E tests src/pages/admin/* — 8 routes, role-based access, CRUD operations ● Medium — Error boundary fallback not verified src/components/ErrorBoundary.tsx — renders fallback UI on crash qmax > generate tests for the critical gaps ✓ 4 test cases created for payment webhooks ✓ 2 test cases created for cart concurrency ✓ 6 Playwright scripts auto-generated Coverage: 34 → 40 test cases (+18%)

Browser auth + secure key management

qmax > /connect /\_/\ / ( o.o ) / > ^ < / /| | Let's connect you to QualityMax! Your auth code: QMAX-8J9EXG Opening browser to authorize... Waiting for authorization... ✓ Connected as dev@acme.io qmax > /keys API Keys Anthropic: ● Set (sk-a•••5xYz) QualityMax: ● Connected (dev@acme.io) qmax > /status qmax-code Status QualityMax: ● Connected Logged in as: dev@acme.io API: https://app.qualitymax.io Mode: standalone (direct API) Active project: #142 Model: sonnet Session tokens: 4,200 in / 890 out Est. cost: $0.0260

Run pytest locally, results sync to platform

qmax > run script 856 locally with base url http://localhost:3000 📥 Downloading script #856 (pytest)... 📝 Writing to /tmp/qmax-test-a7f3/test_script.py 🚀 Running: python3 -m pytest test_script.py -v test_api_health_check PASSED [0.3s] test_list_bridges PASSED [0.8s] test_create_bridge PASSED [1.2s] test_invalid_config FAILED [0.4s] AssertionError: Expected 400, got 422 ✓ 3 passed ✗ 1 failed [2.7s] ✓ Results reported to QualityMax (exec: local_856_1774709037)

Why this exists

Nothing else does this

AI code assistants write tests. Test platforms run them. qmax-code is the first to do both — from your terminal, in natural language.

Capability	qmax-code	AI Assistants (Copilot, Cursor, etc.)	Test Platforms (Stably, Testim, Mabl)
Terminal-native AI agent	✓	—	—
Natural language test commands	✓	code only	—
AI crawl → auto-generate tests	✓	—	record only
Cloud test execution	✓	—	✓
Self-healing scripts	✓	—	basic
k6 performance / load testing	✓	—	—
CI/CD framework export	✓	—	config only
Test case CRUD + management	✓	—	✓
BYOLLM / open source	✓	—	—
Works behind firewalls	✓	✓	—

Pricing

Free to start. Powerful to scale.

qmax-code is included in every QualityMax plan*

Free (BYOLLM)

forever

5 projects, 10 AI crawls/month. Bring your own LLM API key. Full qmax-code access.

Get Started

Starter

€79

/month

25 projects, 100 crawls, cloud execution, k6 perf testing, team seats.

Start Free Trial

Pro

€199