AUTOTEST vs OpenTest.AI

AUTOTEST is an open-source GenAI command-line tool that analyzes web pages and generates automated test cases and Python Selenium scripts. Starting from a base URL, it recursively collects unique internal URLs with breadth-first search to a chosen depth. Generated scenarios can include positive and negative cases for login, registration and forms when test data is supplied. Page analysis combines language-model metadata extraction with Selenium selector functions as a fallback. The project lists Selenium and Playwright support, with Puppeteer described in the development branch. LangChain provides access to major open-source and closed-source language models, with OpenAI, Groq and Anthropic shown as provider examples. Tests can pause for a user to handle a CAPTCHA or another security challenge. Scripts and reports are saved in local folders. Python 3.10.12 and an LLM-provider API key are prerequisites. SQLite caching is available, with a no-cache option to regenerate outputs. The project uses the MIT License.

OpenTest.AI is a self-hosted, open-source project for local-first AI quality checks and portable JSON testing formats. Its six skills include automatic bug fixing, deep quality scans, risk analysis, statistical confidence analysis, smart comparisons, and metrics. The skills run locally without an API key, cloud call, or account, and integrate with clients such as Claude Code, Cursor, Codex CLI, GitHub Copilot Chat, Aider, and Continue.dev. jank_deep scans 12 quality dimensions, while jank_risk follows code blast radius, GenAI smells, user frustration, and business risk. jank_confidence uses test-coverage statistics and estimates the cost and time to close gaps. jank_diff compares APIs, documents, or local files while ignoring timestamps, dynamic IDs, and generated metadata noise. The project analyzes screenshots, network and console logs, and DOM or accessibility trees through 33+ AI testing profiles. It supports bug detection, persona feedback, and prioritized test-case generation. Test cases can be replayed by Playwright, Selenium, Cypress, or a human. The MIT plan is free forever.