AWS Transform Custom · Generally AvailableTransform any code pattern — version upgrades, SDK migrations, framework transitions, and custom transformations. Powered by AI agents that learn, adapt, and scale.
70% of Fortune 500 legacy code is 20+ years old. Manual migration is slow, error-prone, and doesn't scale.
Outdated frameworks, deprecated APIs, end-of-life runtimes. Every delay increases security risk and maintenance cost.
Manual code analysis, dependency mapping, refactoring — each repo takes days or weeks of skilled developer time.
One repo at a time. 500 repos to upgrade? That's years of work with traditional approaches.
Migration expertise lives in a few engineers' heads. When they leave, the knowledge goes with them.
A four-phase workflow: Define → Pilot → Scale → Monitor. The agent improves from every execution through continual learning.
Natural language + docs + samples
Local or 128 parallel on Fargate
Build + test + self-debug
Knowledge items extracted
From low-effort version upgrades to high-complexity cross-language migrations — all driven by natural language
Beyond one-off transformations, AWS Transform can watch your whole portfolio. Connect GitHub, GitLab, or Bitbucket, then run analyses on demand or on a schedule to surface and prioritize tech debt. For findings that have a fix, it opens pull or merge requests with validated code changes for review. Analysis and remediation run in your own AWS account using your credentials — your source code stays under your control.
Scans code for cloud-native maturity gaps and maps findings to AWS modernization pathways. Completes in about 5–30 minutes per repository.
Evaluates whether a system is ready to be safely called by AI agents — covering APIs, identity, state management, human-in-the-loop controls, and observability.
Watch Java 8 code transform into modern Java 21 — character by character.
Pick your repo count, hit Start, and watch the time difference unfold.
Assumes 80% time reduction with AWS Transform (reported by customers). Adjust dev-days to match your environment.
Most AWS Transform agents are free. Custom transformations are billed per agent minute — only when the agent is actively working on your code. AWS-managed .NET transformations now run free up to 50,000 agent minutes a month.
| Agent | Cost |
|---|---|
| Assessment | Free |
| VMware migration | Free |
| Windows modernization | Free |
| Mainframe modernization | Free |
| AWS-managed .NET transformation new | Free* |
The AWS-managed AWS/dotnet-modernization transformation runs at no charge up to 50,000 agent minutes per month — enough for a ~500,000-line .NET Framework port (~25,000 agent minutes) with room to spare. Unused minutes do not roll over.
Standard AWS infrastructure charges apply for resources created during migration (EC2, RDS, etc.).
The migration and modernization agents are available in all AWS commercial regions. AWS Transform custom started in US East (N. Virginia) and has since expanded to eight regions — adding Europe (Frankfurt, London), Asia Pacific (Mumbai, Tokyo, Seoul, Sydney), and Canada (Central).
Region coverage changes often — confirm against the supported regions page before you plan a rollout.
| What's billed | Included? |
|---|---|
| Agent planning & reasoning | Billed |
| Code analysis & modification | Billed |
| Local builds & test runs | Not billed |
| User idle time | Not billed |
| Continual learning | Included |
| Campaign management | Included |
| Continuous modernization analyses new | Billed |
| Continuous modernization remediations new | Billed |
Continuous modernization runs on the same custom agent, so each repo analysed and each remediation is billed in agent minutes. Security analyses are the exception — those are charged by AWS Security Agent, not in agent minutes.
Real migrations executed by AWS Transform Custom — explore the full worklog, code changes, and validation results.
Spring Boot 3.x, Jakarta EE, JUnit 5, records, switch expressions, text blocks, streams, pattern matching — the full Java modernization story.
pytest upgrade, 25+ .format() → f-string conversions, documentation update. Emoji-based algebra game 🍎🐶⚽ — proves multi-language support.
Full cross-language migration using a custom transformation definition. Python classes → TypeScript interfaces, pytest → Jest, readline singleton pattern. Proves custom TDs work for any pattern.
Install the Kiro IDE extension or use the ATX CLI to start transforming today.
Key details from the official AWS Transform FAQ — covering capabilities, security, and scale.
SKILL.md file with optional references/ folder. Create via natural language in interactive mode, save as draft, then publish to your account's registry.
references/ folder. Migration guides, API specs, code samples (max 10MB total). Loaded as needed during execution.atx custom for interactive mode, atx custom def exec for autonomous execution. Supports --trust-all-tools for fully unattended runs. Scriptable into CI/CD pipelines.atx custom def list (registry), atx custom def save-draft (save WIP), atx custom def publish (share with team), atx custom def list-ki (view knowledge items).