ReCAP is a framework for training and evaluating CAPTCHA-capable GUI agents using dynamic tasks, benchmarks, and unified evaluation.
AI & ML interests
None defined yet.
Recent Activity
Organization Card
ReCAP Agent
ReCAP Agent develops models and datasets for CAPTCHA-capable GUI agents.
- Models: GUI agents (ReCAP-8B, ReCAP-32B) trained to solve CAPTCHAs with reasoning-action outputs
- Datasets: Dynamic CAPTCHA datasets with diverse interaction types, including reasoning-action traces and self-correction traces.
Our goal is to provide resources for training and evaluating GUI agents that can reliably operate in CAPTCHA-gated environments.