Weblica: Scalable and Reproducible Training Environments for Visual Web Agents
Apple just open-sourced the training framework that could unlock web agents at scale.
Why it matters
Apple's Weblica framework addresses a critical bottleneck in visual web agent development—reproducible, diverse training environments. This research demonstrates how to scale training data for agents that interact with real-world web interfaces, a capability gap that directly impacts how enterprises will deploy autonomous web automation.
The key facts
10 to knowFramework: Weblica (Web Replica) for constructing reproducible web environments
Technical approach: HTTP-level caching + LLM-based environment synthesis
Problem solved: Scaling training data beyond offline trajectories and simulated environments
Use case: Visual web agents training on diverse, interactive web states
Source: Apple ML Research (academic/infrastructure contribution)
Apple research: HTTP-level caching for reproducible web environment capture
LLM-based environment synthesis for scalable training data generation
Framework solves web diversity problem in visual agent training
Addresses limitations of offline trajectories and simulated environments for RL
Published on machinelearning.apple.com — academic research release
Go to the source
Apple Machine Learningmachinelearning.apple.com
Publisher excerpt: The web is complex, open-ended, and constantly changing, making it challenging to scale training data for visual web agents. Existing data collection attempts remain limited to offline trajectories for supervised fine-tuning or a handful of simulated environments for RL training, thus failing to…
