LaVague is a Large Action Model framework for developers that translates natural-language objectives into automated Selenium or Playwright browser actions. It requires an external OpenAI API key and Python coding to build and run agents locally.
Project overview
LaVague is a Large Action Model framework for developers that translates natural-language objectives into automated Selenium or Playwright browser actions. It requires an external OpenAI API key and Python coding to build and run agents locally.
Project type
AI Agent · Workflow
Use cases
Automation
Deployment
Refer to project documentation
License
Apache-2.0
Best for
Developers and operations teams who need to automate browser workflows using a Python library and are equipped to manage OpenAI API dependencies.
Key capabilities
Takes a user-defined objective and the current state of a web page to output appropriate navigation instructions.
Compiles navigation instructions generated by the World Model into executable action code, such as Selenium or Playwright scripts, and executes them.
A test runner provided for testing and benchmarking the performance of LaVague agents.
A tool for estimating token usage and the associated costs of running agents.
An optional, interactive Gradio interface for managing and interacting with agents.
A tool tailored for QA engineers to automate test writing by turning Gherkin specifications into executable tests.
Limitations and risks
The framework collects user data telemetry by default, including objectives and HTML code chunks. This poses a privacy risk if personal information is included, though telemetry can be configured as optional.
Running agents requires an external OpenAI API, meaning a paid service is necessary to operate the framework's default LLM functionality.
Getting started
Install the framework using a single pip command, set your OpenAI API key, and run a Python script to instantiate a WebAgent and execute an objective.
Evidence and sources
GitHub project description: Large Action Model framework to develop AI Web Agents
README: LaVague is an open-source framework designed for developers who want to create AI Web Agents to automate processes for their end users.
README: Note that not all drivers support all agent features:
README: A test runner for testing and benchmarking the performance of LaVague
README: LaVague QA is a tool tailored for QA engineers leveraging our framework. It allows you to automate test writing by turning Gherkin specs into easy-to-integrate tests.