A curated list of AI tools and resources for developers, see the AI Resources .

LaVague

LaVague is an open-source framework for building Web Agents that translate high-level objectives into executable browser actions.

Detailed Introduction

LaVague is an open-source framework for developers to build Web Agents that perform multi-step tasks on the web. It composes a World Model, an Action Engine and drivers (Selenium, Playwright, or Chrome extension) to compile high-level objectives into executable actions and run them automatically, enabling browser automation and task execution workflows.

Main Features

  • Multi-driver support: compatible with Selenium, Playwright and a Chrome extension for different execution environments.
  • Modular architecture: World Model, Action Engine and configurable contexts support extensibility and testing.
  • QA tooling and examples: includes LaVague QA, demos and quick-start guides to accelerate development and verification.
  • Observability: built-in logging, token counting and telemetry controls for cost and behavior monitoring.

Use Cases

Suitable for scenarios that require orchestrated browser automation, such as automated testing, web data collection, user-flow automation, and integrating web-based toolchains. Teams focused on privacy or local operation can leverage local drivers and self-hosted deployments.

Technical Features

  • Engineering-oriented: pluginable modules and configurations for CI/CD and testing pipeline integration.
  • Composable execution pipeline: converts high-level objectives into action sequences executed by drivers for complex task orchestration.
  • Open-source licensing: Apache-2.0 licensed for auditability and enterprise integration.
  • Comprehensive docs and examples: documentation site and Colab notebooks for reproducibility and onboarding.
LaVague
Resource Info
🎨 Multimodal ⚙️ Automation 🌱 Open Source