testzeus hercules

Report Abuse

Basic Information

Hercules is an open-source, AI-driven testing agent that turns Gherkin feature files into fully automated end-to-end tests without requiring users to write code. It orchestrates browser interactions using Playwright, plans and decomposes tasks with LLM-based agents, and targets complex web platforms such as Salesforce. Hercules is designed to run locally, in Docker containers, or inside CI/CD pipelines and can connect to remote browser instances via CDP. The project bundles reporting and proof artifacts including JUnit XML, HTML reports, screenshots, videos, network logs and detailed chat_messages logs for introspection. It supports configurable LLM backends, environment-driven configuration, interactive instruction execution mode for RPA and debugging, and is extensible by adding custom tool modules and LLM configurations.

Links

Categorization

App Details

Features
Gherkin-in, results-out workflow that converts readable feature files into executable tests. Multi-agent architecture with a planner agent and a browser navigation agent that use LLMs for planning and tool invocation. Tools library of sensing and action primitives (openurl, click, enter_text, DOM distillation utilities) to interact with web pages. Playwright integration for browser control, mobile device emulation, video and screenshot capture, and network logs. Docker image and CLI with environment variable configuration for CI/CD. Extensible additional tools directory and agents_llm_config.json for custom LLMs. Built-in accessibility checks, security testing integration via Nuclei, multilingual support, and tracing/playwright trace export.
Use Cases
Hercules simplifies and democratizes end-to-end testing by enabling non-developers to author tests in Gherkin while leveraging LLMs to handle planning and browser interactions. It reduces maintenance through autonomous, adaptive execution and provides robust evidence for failures with videos, screenshots, network logs and structured reports suitable for CI. Teams can test complex UIs, run accessibility and security checks alongside functional tests, emulate mobile devices, and scale runs using Docker or remote browser providers. The agentic design makes it easier to extend behavior with new tools and custom LLM configurations, and an interactive mode supports RPA-style debugging and exploratory runs. Token and model usage are configurable to manage LLM costs.

Please fill the required fields*