Hypermantic has launched Hopper, an AI-powered development environment for mainframe systems. The tool aims to bridge the gap between legacy COBOL-based infrastructure and modern AI capabilities, targeting the aging but critical mainframe systems that still process banking, insurance, and government transactions worldwide. Most of these systems run on IBM z/OS with decades-old code that remains reliable and secure.

Hopper combines three essential components: a real TN3270 terminal, mainframe-aware panels for datasets and jobs, and an AI agent that operates across z/OS surfaces. Unlike generic chatbots, Hopper works directly within the mainframe environment, preserving system fidelity while enabling automation of repetitive tasks like debugging and code changes. The tool can interpret JES return codes, locate source errors, and suggest fixes autonomously.

The tool addresses a real pain point: mainframe development remains stuck in decades-old workflows with TN3270 terminals, ISPF panels, and specialized conventions. Hopper offers concrete solutions like faster job debugging and automated documentation without abstracting away the underlying system complexity. Users can request free access to try Hopper on Hypercubic's mainframe environment immediately.