Amazon Web Services launched AWS Transform custom, which is a new agent to modernize the custom code that got you into your technical debt pickle in the first place.
The launch, which AWS kicked off at re:Invent 2025 with a stunt that included blowing up a server rack, is the latest in a portfolio of Transform services.
- AWS, Google Cloud engineer interconnect between their clouds
- AWS Marketplace adds solutions-based offers, Agent Mode
In a blog post, AWS combines pre-built transformations to upgrade anything in Java, Node.js, and Python with the ability to define custom work.
How do you crush #techdebt? @awscloud says #Transform with agentic AI to eliminate the lag and modernize without the pain and strain. #AWSreInvent pic.twitter.com/rwim14qidK
— Liz Miller (@lizkmiller) December 1, 2025
The goal is to use the agent to free up developers.AWS Transform custom includes:
- Command line and web interfaces to define transformations through natural language and execute them on local codebases.
- AWS Transform custom can operate autonomously.
- The web interfaces provides tracking for teams and transformation progress.
- The system supports runtime upgrades without the need for additional information.
- AWS Transform custom learns enterprise code patterns that have evolved over time.
- DevOps can configure AWS Transform custom with integration and continuous delivery tools.
AWS also announced the following AWS Transform updates.
- AWS followed up the general availability of AWS Transform for .NET with full stack Windows modernization. AWS Transform will cover application, UI, database, and deployment layers and map Windows stacks to AWS services.
- AWS Transform for mainframe adds capabilities called Reimagine, which uses AI to revamp an enterprise architecture using business logic, patterns, data lineage and legacy source code. With the service, customers can transform mainframe applications to microservices and modern architecture. AWS Transform for mainframe also includes automate test plan generation, collection scripts and AI tools to accelerate testing timelines.
Holger Mueller, an analyst at Constellation Research, said:
"AI has tremendous power when transforming and translating language, and code is nothing but a (very structured) language. As such it has a lot of promise for code renovation and modernization. Good to see AWS tackling this, but it is kind of skimming the low hanging fruit by modernizing 5-10 year old code assets. Staying away from 10-29+ year old code assets, these require code translation from the legacy to a modern programming language. And that is where the prize for CxOs really is."
