Hypercubic: AI tools for mainframe and COBOL systems

What it does

Hopper is a desktop application that combines a TN3270 terminal emulator with an AI agent for mainframe and COBOL development. It runs on Windows, macOS, and Linux. The agent can navigate ISPF panels, write JCL code, submit and monitor batch jobs, debug failed jobs by parsing error messages, query VSAM records, and read COBOL source files. All operations happen on the user's laptop; nothing is installed on the mainframe itself. The tool connects to z/OS using standard TN3270 protocol for terminal access and optional FTP for datasets and job queues.

Who it is for

Hopper targets z/OS engineers and mainframe developers. The product is aimed at teams maintaining legacy COBOL systems and batch job infrastructure. It serves organizations that need faster development cycles and better debugging workflows for mainframe applications.

Pricing

Hopper offers a free Hobby tier that includes all core features, no credit card required, and works on macOS, Windows, and Linux. An Enterprise tier with custom pricing includes SAML SSO, MCP Server access, admin controls, organization-wide privacy settings, on-premises or VPC deployment, and priority support. Customers must contact sales for Enterprise pricing.

How it stands out

Most mainframe development still relies on decades-old terminal interfaces and manual processes. Hopper adds an AI agent layer that understands mainframe-specific concepts like column-strict JCL syntax, ISPF panel navigation, and JESMSGLG error decoding. The agent can chain multiple mainframe operations together from a single natural language prompt. The product requires no mainframe-side installation or system programmer involvement, reducing deployment friction. It provides a modern graphical interface while preserving full compatibility with the legacy TN3270 protocol.

What a founder should check

Prospective competitors should investigate how many z/OS shops would actually adopt a new development tool versus sticking with established workflows. Switching costs are potentially high—teams have muscle memory with ISPF, SDSF, and existing automation scripts that would need to be replicated or abandoned. The real moat may be whether the AI agent's mainframe knowledge compounds over time or whether incumbents like IBM quickly add similar capabilities to their existing tools and products. Also verify whether enterprises view the free Hobby tier as a viable go-to-market funnel or whether mainframe shops require immediate enterprise features and pricing that make the freemium model ineffective at generating revenue.

Thinking of building something like this?

Every launch here is a competitor to somebody's idea. If yours is close, check it against the market before you build: the Full Check names the rivals, the prices and the gaps.