CAPTCHA Solving for Native GUI Agents: Automated Reasoning-Action Data Generation and Self-Corrective Training

Yuxi Chen , Haoyu Zhai , Chenkai Wang , Rui Yang , Lingming Zhang , Gang Wang , Huan Zhang

🏛 Institutions: UIUC
📅 Date: March 23, 2026
📑 Publisher: arXiv
💻 Env: General GUI
🔑 Keywords: CAPTCHA reasoning-action data generation self-correction native GUI agent ReCAP

TLDR

ReCAP is a native GUI agent specialized for interactive CAPTCHA solving. It builds a seven-type CAPTCHA environment, generates large-scale reasoning-action trajectories plus self-correction data from failed attempts, and improves success from about 30% to 80% without sacrificing general GUI-agent performance.

Open paper Report issue