Welcome to LaVague

A Large Action Model framework for developing AI Web Agents

LaVague is an open-source Large Action Model framework which aims to leverage advanced AI techniques (RAG, Few-shot learning, Chain of Thought) to develop effective AI Web Agents.

Our web agents take an objective, such as "Print installation steps for Hugging Face's Diffusers library Thought:" and performs the required actions to achieve this goal by leveraging our two core components:

• A World Model to break down an objective into instructions for step-by-step web actions
• An Action Engine which uses to “compile” these instructions into automation code, by leveraging Selenium or Playwright & execute them

You can download the LaVague PyPi package with:

pip install lavague

You can then leverage our library to automate web actions based on natural language objectives:

from lavague.retrievers import OpsmSplitRetriever
from lavague.defaults import DefaultEmbedder, DefaultLLM, default_get_selenium_driver
from lavague.action_engine import ActionEngine
from lavague.world_model import GPTWorldModel
from lavague.agents import WebAgent
from lavague.web_utils import resize_driver

driver = default_get_selenium_driver()
action_engine = ActionEngine(DefaultLLM(), OpsmSplitRetriever(DefaultEmbedder(), top_k=3))
world_model = GPTWorldModel()

agent = WebAgent(driver, action_engine, world_model)
agent.get("https://huggingface.co/docs")
agent.run("Go on the quicktour of PEFT")

For more information on this example and how to use LaVague, see our quick-tour.

Note, these examples use our default OpenAI API configuration and you will need to set the OPENAI_API_KEY variable in your local environment with a valid API key for these to work.

For an end-to-end example of LaVague in a Google Colab, see our quick-tour notebook

We would love your help and support on our quest to build a robust and reliable Large Action Model for web automation.

To avoid having multiple people working on the same things & being unable to merge your work, we have outlined the following contribution process:

1. 📢 We outline tasks on our backlog: we recommend you check out issues with the help-wanted labels & good first issue labels
2. 🙋‍♀️ If you are interested in working on one of these tasks, comment on the issue!
3. 🤝 We will discuss with you and assign you the task with a community assigned label
4. 💬 We will then be available to discuss this task with you
5. ⬆️ You should submit your work as a PR
6. ✅ We will review & merge your code or request changes/give feedback

Please check out our contributing guide for a more detailed guide.

If you want to ask questions, contribute, or have proposals, please come on our Discord to chat!

TO keep up to date with our project backlog here.

Note, this project executes LLM-generated code using exec. This is not considered a safe practice. We therefore recommend taking extra care when using LaVague (such as running LaVague in a sandboxed environment)!

By default LaVague records some basic anonymous values to help us gather data to build better agents and Large Action Models:

• Version of LaVague installed
• Code generated for each web action step
• LLM used (i.e GPT4)
• Randomly generated anonymous user ID
• Whether you are using a CLI command or our library directly
• The URL you performed an action on
• Whether the action failed or succeeded
• Error message, where relevant
• The source nodes (chunks of HTML code retrieved from the web page to perform this action)

If you want to turn off telemetry, you can set your TELEMETRY_VAR environment variable to NONE in your working environment.