Software engineering. Legacy system modernization and migration. Native Rust applications. Evidence you can inspect.
Genyra LLC engineers native applications, enterprise platforms and systems integrations. We investigate existing software—its business rules, data and dependencies—and build the next implementation around behaviour that can be tested: COBOL applications reconstructed in Rust and verified byte for byte, and native desktop software measured against the original it replaces.
Genyra.com · Engineering Lab on Genyra.com · Engineering Lab index · Discuss a project
These four application repositories are forks of upstream open-source projects. Genyra's contribution is the Rust implementation, migration tooling, verification and documentation described below. Original authorship and licences remain attributed in each repository.
Genyra rebuilt 13 COBOL programs as a standalone Rust banking-and-payroll demonstration, including exact decimal arithmetic and packed-decimal records. Published verification covers 90 simulated days, six nodes and 2,174 byte-for-byte file comparisons, plus the employee payroll pipeline.
Evidence and scope · Lab page · Upstream: Albert Mendes / AKD APP BREWING's Legacy Ledger.
Genyra rebuilt the COBOL simulator's reservations, stays, services, loyalty and billing workflows in native Rust, with a JSON command interface and persistent snapshots. 155 scenarios and 7,351 byte-for-byte table-record comparisons are recorded with zero mismatches.
Evidence and scope · Lab page · Upstream: Alejandro Blanco's COBOL hotel simulator.
3. COBCALC
Genyra rebuilt the expression evaluator in Rust with an explicit decimal contract, offline dependencies and reproducible checks. Evidence separates 2,000 compatibility cases against the original floating-point COBOL program from 1,000 exact-decimal reference cases. The two comparisons establish different things.
Evidence and scope · Lab page · Upstream: Manny Juan's COBCALC.
Linux alpha. Genyra replaced the Python collector with Rust for network visibility, application controls and investigation tools. The native desktop and Windows/macOS work remain under validation; full application parity is still in progress.
The 4 November 2026 Linux measurements cover background collectors and the revised native application. Background mean RSS was 109.94 → 44.89 MiB; complete-app mean summed RSS was 1,279.49 → 171.89 MiB, with PSS 755.24 → 113.77 MiB. These five-minute workloads used differing capture policies and animation cadences. Native application RSS still grew during its run; neither sample establishes long-term stability or full parity. The earlier higher-native-RSS result is retained in the resource guide.
Measurements and limits · Lab page · Upstream: Snitch by aiXis Studio.
The Engineering Lab links to source provenance, runnable demonstrations, comparison methods and recorded results. These are scoped engineering demonstrations; broader compatibility and production readiness require their own evaluation.
For a larger legacy system migration, software modernization or engineering project, talk to Genyra.