Raiken

AI-powered E2E test generation, execution, and repair.

Raiken is a developer-centric CLI tool that uses AI to automatically generate, run, and fix Playwright end-to-end tests. It understands your codebase through semantic analysis and generates tests with accurate selectors from your live application.

Features

• Natural Language Test Generation — Describe what you want to test, get working Playwright code
• Smart Codebase Understanding — AST parsing + embeddings understand your app structure
• Live DOM Capture — Captures real page elements for accurate, stable selectors
• Visual Dashboard — Modern UI for generating, viewing, and managing tests
• Self-Monitoring Tests — Automatically detects stale tests and suggests repairs

Installation

npm install -g raiken

Requires Node.js 18+

Quick Start

1. Initialize in your project

cd your-project
raiken init

This creates a raiken.config.json and sets up the test directory.

2. Set your API key

Raiken uses OpenRouter to access AI models (Claude, GPT-4, etc.).

export OPENROUTER_API_KEY=your_api_key_here

Or add it to a .env file in your project root:

OPENROUTER_API_KEY=sk-or-v1-your-key-here

Get a free API key at openrouter.ai/keys

3. Start the dashboard

raiken start

Open http://localhost:7101 in your browser.

4. Generate your first test

In the dashboard chat, describe what you want to test:

"Test the login flow at localhost:3000 - enter email and password, click submit, verify redirect to dashboard"

Raiken will:

1. Analyze your codebase for context
2. Capture the live DOM from your running app
3. Generate a Playwright test with real selectors
4. Save it to your test directory

Commands

| Command | Description | |---------|-------------| | raiken init | Initialize Raiken in your project | | raiken start | Start the server and dashboard on port 7101 | | raiken start -p 8080 | Start on a custom port |

Configuration

Create raiken.config.json in your project root:

{
  "testDirectory": "e2e",
  "baseUrl": "http://localhost:3000",
  "ai": {
    "model": "anthropic/claude-sonnet-4.5"
  }
}

Configuration Options

| Option | Type | Default | Description | |--------|------|---------|-------------| | testDirectory | string | "e2e" | Where to save generated tests | | baseUrl | string | "http://localhost:3000" | Your app's development URL | | ai.model | string | "anthropic/claude-sonnet-4.5" | OpenRouter model to use |

Supported AI Models

Any model available on OpenRouter, including:

• anthropic/claude-sonnet-4.5 (recommended)
• anthropic/claude-opus-4
• openai/gpt-4o
• google/gemini-2.0-flash
• meta-llama/llama-3.3-70b

Environment Variables

| Variable | Required | Description | |----------|----------|-------------| | OPENROUTER_API_KEY | Yes | Your OpenRouter API key | | OPENROUTER_MODEL | No | Override the AI model | | OPENROUTER_BASE_URL | No | Custom API endpoint |

How It Works

┌─────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│  Your Codebase  │────▶│  Raiken Analyzer │────▶│  Code Graph DB  │
└─────────────────┘     └──────────────────┘     └─────────────────┘
                                                         │
┌─────────────────┐     ┌──────────────────┐             │
│  Your Running   │────▶│  DOM Capture     │             │
│  Application    │     │  (Playwright)    │             │
└─────────────────┘     └──────────────────┘             │
                                │                        │
                                ▼                        ▼
                       ┌──────────────────────────────────┐
                       │         AI Test Generator        │
                       │    (Claude/GPT-4 via OpenRouter) │
                       └──────────────────────────────────┘
                                        │
                                        ▼
                       ┌──────────────────────────────────┐
                       │     Generated Playwright Test    │
                       │         saved to e2e/            │
                       └──────────────────────────────────┘

1. Codebase Analysis — Raiken parses your source files, extracts functions, components, and routes, then creates embeddings for semantic search.

2. DOM Capture — When you describe a test, Raiken launches a headless browser, navigates to your app, and captures all interactive elements with their selectors.

3. Context Assembly — Relevant code snippets + live DOM elements are assembled into a rich context for the AI.

4. Test Generation — The AI generates a Playwright test using real selectors from your application.

5. Test Execution — Run generated tests with Playwright:

   npx playwright test

Project Structure

After initialization, Raiken creates:

your-project/
├── raiken.config.json    # Configuration
├── .raiken/
│   └── raiken.db         # Local SQLite database (code graph, embeddings)
└── e2e/                  # Generated tests (configurable)
    └── login.spec.ts

Running Generated Tests

Raiken generates standard Playwright tests. Run them with:

# Run all tests
npx playwright test

# Run specific test
npx playwright test e2e/login.spec.ts

# Run with UI
npx playwright test --ui

If you don't have Playwright installed:

npm init playwright@latest

Troubleshooting

"OPENROUTER_API_KEY not configured"

Set your API key:

export OPENROUTER_API_KEY=sk-or-v1-xxxxx

Or add it to .env in your project root:

OPENROUTER_API_KEY=sk-or-v1-xxxxx

Dashboard not loading

Make sure port 7101 is available, or use a different port:

raiken start -p 8080

DOM capture fails

Ensure your application is running at the URL you specify in your test prompt (e.g., "test login at localhost:3000").

License

MIT © Raiken Foundation

Links

• GitHub Repository
• Report Issues
• OpenRouter API