Amazon Q Developer review

What it is

Amazon Q Developer is AWS's coding assistant, available in IDEs, the CLI and the AWS console. It competes less on general code generation and more on knowing AWS in depth.

Where it shines

Cloud specificity. Ask why a Lambda times out when writing to DynamoDB, or to write an IAM policy granting exactly one action on one bucket, and it answers with the accuracy of something trained on the full AWS documentation set and service behaviour. General assistants hallucinate service limits and parameter names; Q usually does not.

The Java transformation feature is the sleeper hit. Point it at a legacy Java codebase and it upgrades across major versions, updating dependencies, fixing deprecated APIs and running tests. Enterprises with a decade of Java debt report migrations completing in days that were previously quoted in quarters. Nothing else in this category does anything comparable.

Security scanning with in-place suggested fixes rounds out a package aimed squarely at large cloud engineering organisations.

Where it falls short

Outside AWS, it is unremarkable. Front-end work, general algorithms, non-AWS backend code — all handled competently but below what Copilot, Cursor or Claude Code produce.

The experience is also fragmented. The console assistant, the IDE plugin and the CLI behave differently and have different capabilities, which is confusing. Documentation assumes you already know which one you are using.

Pricing in plain English

Free tier with monthly limits that suit individual use. $19 per user per month for Pro, which unlocks transformation jobs and admin controls, and $39 per user per month for Team if you need the full codebase agent and unlimited security scans.

Who should use it

AWS-centric engineering teams, especially those maintaining large Java estates. Run it alongside a general-purpose assistant rather than as a replacement.

A note on adoption

The most successful rollouts start with the transformation features rather than the chat. Pick one legacy Java service, run the upgrade, measure the hours saved against the previous quote, and use that number to justify wider licensing. Teams that lead with general chat comparisons instead tend to conclude it is worse than Copilot, which is true in that narrow test and misses where the product actually earns its money.