Kodesage closes $6.6M seed to scale AI-powered legacy codebase understanding across regulated industries

Many large enterprises in sectors such as banking, insurance, energy and the public sector are still running critical operations on software written decades ago. Modernising this software is no longer optional - regulatory pressure, talent shortages, and technical debt make it urgent. However, the process is slow and costly, partly because the first obstacle is simply understanding what the existing code does.

Kodesage, founded in 2024 by Gergely Dombi, Miklós Szurdi and György Szilágyi, has developed an on-premises AI platform that addresses this issue by helping to understand the code. The platform reads legacy source code in various languages, including COBOL, PL/SQL, Oracle Forms, PowerBuilder, and RPG, alongside existing documentation, databases, and ticket systems. The platform then builds a continuously updated knowledge layer from the information it finds. This foundation then powers auto-generated documentation, context-aware code conversion, test creation and AI-assisted production support. The system runs entirely inside customer-controlled infrastructure, whether on-premises, in a private cloud or in a fully air-gapped environment, ensuring that source code and business-critical data remain within the organisation.

In June 2026, Kodesage closed a $6.6 million seed round led by VentureFriends, with existing investor Portfolion returning alongside angel investors. This follows a €2.3 million pre-seed round closed in January 2025, bringing the total funding to approximately $8.9 million. The capital will support expansion into the US and European markets and continued product development.

Sources: Kodesage | vestbee

Founders: Gergely Dombi, Miklos Szurdi, Gyorgy Szilagyi