<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> Skiritai is an AI-powered test automation framework designed to accelerate web application testing. It addresses the time-consuming process of creating and executing test scripts by using an AI agent to explore application paths, generate replayable scripts, and then execute these scripts at significantly higher speeds. This benefits engineers and QA teams by drastically reducing the time spent on writing and running regression tests.

How It Works

Skiritai employs an "Explore -> Replay" loop. Initially, an AI agent navigates the target application, analyzes UI elements, and determines optimal action sequences, generating replayable scripts. In subsequent runs, these generated scripts execute directly, bypassing AI inference and achieving up to 30x faster execution (e.g., 74s to 2.5s). This approach combines AI-driven discovery with high-performance, deterministic script execution.

Quick Start & Requirements

• Installation: pip install skiritai and playwright install chromium.
• Prerequisites: Python 3.11+, an API key for a supported LLM (OpenAI, Anthropic, Qwen) configured via .env file (e.g., OPENAI_API_KEY, LLM_MODEL).
• Running Examples: Use the CLI command skiritai run <case_dir> or execute Python scripts directly (e.g., python examples/beginner/baidu_search/02_flow/demo.py).
• Documentation: Examples are provided for various use cases, including beginner tutorials and advanced scenarios.

Highlighted Details

• Explore → Replay Loop: AI scouts and generates scripts on the first run; subsequent runs execute these scripts instantly.
• 30x Performance: Replay mode skips AI inference, drastically reducing execution time.
• Flexible APIs: Supports both Python-native BaseCase and flow() functional APIs, as well as declarative YAML test cases.
• LLM Agnostic: Compatible with OpenAI, Anthropic, Qwen, and other compatible APIs.
• Visual Reports: Generates standalone HTML reports with screenshots, assertions, and step details.
• CLI & Optional Web UI: Provides command-line tools for running, listing, and serving tests, with an optional FastAPI-based web UI.

Maintenance & Community

The project is primarily maintained by Joe Shen. Contributions are welcomed via pull requests. No specific community channels (like Discord or Slack) are listed in the README.

Licensing & Compatibility

The project is released under the MIT License, which permits commercial use and integration with closed-source projects.

Limitations & Caveats

Currently, Skiritai focuses on web browser automation using Playwright. Planned extensions include mobile (iOS/Android) and API testing. The AI's current perception relies on DOM analysis and CSS selectors; future development aims to incorporate visual perception using multimodal models.