lavague.ai
LaVague.ai is the team behind LaVague, an open-source framework for building AI Web Agents that automate repetitive web and business processes. Its stated goal is to democratize AI Web Agents and automate mundane, mechanical work while allowing people to focus on reasoning and planning.
At a Glance
- Software developers and AI builders
- QA engineers and software-testing teams
- Enterprises adopting AI agents
- Organizations automating CRM, ERP and other SaaS workflows
- +1 more
AI Tools by lavague.ai
(1)
LaVague
Open Source Web Agent Framework
Discussions
No discussions yet
Be the first to start a discussion about lavague.ai
Latest News
LaVague repository's latest displayed main-branch update; project shows more than 700 commits and continued work on engines, integrations and Chrome tooling.
LaVague QA CLI announced for automation engineers, turning Gherkin specifications into automated tests.
LaVague open-source Web Agent project launched after a hackathon and reached number one on Hacker News.
Products & Services
Open-source Large Action Model framework for building AI Web Agents. A World Model converts an objective and current web-page state into instructions, while an Action Engine compiles and executes actions through browser drivers such as Selenium or Playwright.
QA-focused tooling built on LaVague that turns Gherkin specifications into easy-to-integrate automated tests and is intended to make web-test authoring substantially more efficient.
Market Position
LaVague differentiates itself as an open-source, developer-oriented Large Action Model framework rather than a closed browser-automation application: it exposes the World Model and Action Engine, supports multiple browser drivers, can use local or remote models, and offers an Apache-2.0 core. Tracxn identifies Hugging Face, Replicate and Flower among its tracked competitors, while LaVague's practical differentiation is its focus on natural-language-to-browser actions, QA automation and customizable/self-hosted deployments.
Leadership
Founders
Daniel Huynh
AI engineer by training; previously CEO of Mithril Security, a Paris-based privacy and security startup. He created LaVague as a hackathon project and led its open-source launch.
Executive Team
Daniel Huynh
CEO
AI engineer and former CEO of Mithril Security; he originated LaVague as a hackathon project and led its open-source development and launch.
Founding Story
LaVague began in early March 2024 as Daniel Huynh's side project for an LLM function-calling hackathon. He wanted to automate mechanical web tasks by generating Selenium code from natural-language instructions; after the demo worked, the team open-sourced it, reached number one on Hacker News, and decided to allocate Mithril Security resources to make AI-powered automation broadly accessible.
Business Model
Revenue Model
Open-core model: the core framework is open source, while planned or offered enterprise features such as security, compliance, auditability and scalability are sold to enterprise customers. The website also offers customized enterprise support and describes a hosted solution as part of the roadmap. Users separately bear the cost of underlying LLM API calls.
Pricing Tiers
LaVague is distributed under Apache License 2.0; operating costs depend on the selected LLM, objective complexity and site being automated.
Customized support to help companies explore and adopt AI-based Agents; the site directs prospective customers to contact the team.
Target Markets
- Software developers and AI builders
- QA engineers and software-testing teams
- Enterprises adopting AI agents
- Organizations automating CRM, ERP and other SaaS workflows
- Privacy-sensitive users needing local or VPC-deployed models
- Web QA and automated test generation from Gherkin specifications
- Process automation such as navigation, data entry and orchestration across SaaS tools including CRM and ERP systems
- Fine-grained information retrieval from SaaS applications
- Automating repetitive browser tasks and web research
- Developer-built AI Web Agents and web automation pipelines