Nitro Test Drive

Zero tests written. Full proof delivered.

6 AI agents discover, test, validate, and self-heal your web app — automatically. Point it at a URL, get a tamper-proof audit report. $99/mo.

Website · Docs · Pricing

What It Does

You give it a URL and credentials. 6 coordinated AI agents:

1. Discover every testable feature in your app (zero manual mapping)
2. Generate test workflows for valid AND invalid paths
3. Execute them with 8-strategy self-healing selectors
4. Validate results with root-cause analysis on failures
5. Repair broken selectors for the next run
6. Report everything in a legal-grade, tamper-proof audit

You write zero test code. You maintain zero selectors.

Quick Start

npm install -g @nitro-ai/test-drive
ntd activate YOUR-LICENSE-KEY

Create ntd.config.json in your project root:

{
    "baseUrl": "https://yourapp.com",
    "auth": {
        "loginUrl": "/login",
        "email": "[email protected]",
        "password": "ENV:TEST_PASSWORD"
    }
}

Run:

export GEMINI_API_KEY=your-gemini-key
ntd agent --cycles 1

That's it. Come back in 3 minutes to a full report.

The 6-Agent Pipeline

┌────────────┐    ┌────────────┐    ┌────────────┐
│ Orchestrator│───▶│ Calibrator │───▶│   Seeker   │
│  (cycle mgr)│    │(classifier)│    │(discovery) │
└────────────┘    └────────────┘    └────────────┘
                                          │
┌────────────┐    ┌────────────┐    ┌────────────┐
│   Repair   │◀───│ Validator  │◀───│  Executor  │
│(self-heal) │    │ (scoring)  │    │(AI testing)│
└────────────┘    └────────────┘    └────────────┘

Each agent operates in a one-way diode chain — can only read its input handoff and produce its output. No backflow. No tampering. Cryptographic signatures at every step.

| Agent | Role | | ---------------- | -------------------------------------------------------------- | | Orchestrator | Manages project memory, cycle state, coordinates all agents | | Calibrator | Classifies discovered features, identifies skip-negatives | | Seeker | Logs in, navigates every page, discovers all testable features | | Executor | AI-generates test workflows, executes valid + invalid paths | | Validator | Scores results, diagnoses failures with root-cause analysis | | Repair | Patches broken selectors, updates silo for next cycle |

Key Systems

🛡️ Sentinel (Security Scanner)

Every run scans for injection patterns, exfiltration attempts, malicious overlays, and DOM tampering. Compromised features are quarantined automatically.

🛑 LOGIC Protocol (Failsafe)

Hardwired anomaly detection. If test counts drop unexpectedly, timing patterns shift, or signatures don't match — the pipeline halts. Human-only unlock required.

⚖️ HAT (Human Autonomy Threshold)

Every action is risk-scored 0-100. Actions above YOUR threshold require approval before execution. The AI can never change its own limits. You stay in control.

📡 Pre-Flight Drift Detection

Before each run, scans the live app against last-known state. Detects new features, removed elements, and structural changes — amends agent instructions automatically.

🩹 Self-Healing Selectors

8 fallback strategies when selectors break. 500+ heals per run is typical. Zero maintenance overhead for selector drift.

CLI Reference

Usage: ntd <command> [options]

Primary:
  agent [--cycles N]   Run the 6-agent AI pipeline (default: 1 cycle)
  agent --no-gate      Skip approval gates (CI mode)

Setup:
  activate <key>       Activate license key
  license              Show current license info
  deactivate           Remove license from this machine

Legacy Modes:
  run                  Quick mode — self-heal + flaky + performance
  drive                Full mode — orchestrated 3-phase pipeline

Analysis:
  report               Regenerate audit report
  status               Pipeline + LOGIC Protocol status

Options:
  --cycles N           Number of test cycles (default: 1)
  --no-gate            Skip human approval gates
  --version, -v        Show version
  --help, -h           Show help

Configuration

ntd.config.json:

{
    "baseUrl": "https://yourapp.com",
    "auth": {
        "loginUrl": "/login",
        "email": "[email protected]",
        "password": "ENV:TEST_PASSWORD"
    },
    "autonomy": {
        "humanThreshold": 60,
        "mode": "supervised"
    },
    "preflight": {
        "enabled": true
    },
    "sentinel": {
        "allowedDomains": ["yourapp.com", "cdn.yourapp.com"]
    }
}

What You Get vs. The Competition

| Capability | Playwright | Cypress | Mabl ($500+/mo) | QA Wolf ($4K+/mo) | Test Drive ($99/mo) | | ---------------------- | ---------- | ------- | --------------- | ----------------- | ----------------------- | | Zero tests to write | ✗ | ✗ | ✗ | ✗ (humans) | ✅ AI discovers | | AI agent pipeline | ✗ | ✗ | ✗ | ✗ | ✅ 6 agents | | Self-healing | ✗ | ✗ | ✓ | ✓ | ✅ 8 strategies | | Security scanning | ✗ | ✗ | ✗ | ✗ | ✅ Sentinel | | Tamper-proof results | ✗ | ✗ | ✗ | ✗ | ✅ SHA-256 | | Human control failsafe | ✗ | ✗ | ✗ | ✗ | ✅ HAT + LOGIC | | Runs on your machine | ✓ | ✗ | ✗ | ✗ | ✅ 100% local |

Pricing

$99/mo — everything included. No tiers. No feature gates. No "contact sales."

• Full 6-agent AI pipeline
• Unlimited projects
• Unlimited team members
• All security, audit, and compliance features
• Cancel anytime

Subscribe →

Requirements

• Node.js 18+
• Playwright (peer dependency)
• Gemini API key (for AI agents — get one free)

License

Commercial license. See LICENSE for terms.

© 2026 NITRO AI LLC — All rights reserved.

Built in Wisconsin. Woman-owned.