Amazon launched Q Developer in April 2024 after realizing CodeWhisperer was getting destroyed by GitHub Copilot. So they rebranded it, jacked up the price to $19/month, and called it "the most capable AI assistant." That's marketing bullshit, but it does have one thing going for it - it actually understands AWS services better than any other coding tool.
The Three Things Q Developer Actually Does Well
Look, I've been using this thing for months and it's frustrating as hell most of the time. But there are exactly three specific areas where it doesn't completely suck:
AWS Infrastructure Code: When you're writing CloudFormation or CDK, Q Developer knows what the fuck it's doing. It suggests proper resource configurations, understands IAM policies, and won't suggest some random garbage that throws errors when you deploy. I've seen it catch S3 bucket policy mistakes that would've taken down prod.
Java Legacy Migration: The Java transformation thing actually works pretty well. We used it to upgrade a bunch of old Spring Boot apps from Java 8 to 17, and it handled most of the dependency hell automatically. Saved us weeks of manual work. Still had to fix some edge cases, but way better than doing it by hand.
Security Scanning That Finds Real Bugs: The security scanner caught SQL injection vulnerabilities in our Node.js code that SonarQube missed. Found hardcoded AWS credentials in test files. It's not perfect but it finds actual problems, not just style issues.
Why It Pisses Me Off Most Days
Slow as Hell: Every suggestion takes 3-4 seconds. You know what kills productivity? Waiting for an AI tool to think while you're trying to code. Cursor gives you suggestions instantly. Q Developer makes you wait like you're on dial-up.
Pushes AWS Everything: Ask it to help with a simple database query and it suggests DynamoDB. Need a message queue? Obviously you want SQS. Sometimes a PostgreSQL database is fine, but Q Developer acts like AWS sales.
General Programming Sucks: Try to get help with React hooks and it gives you code from 2019. Ask about Docker optimization and it suggests basic shit you already know. It's only good at AWS-specific stuff.
The Real Cost Analysis
$19/month is almost double what GitHub Copilot costs, and for what? IP indemnity that your lawyers love but doesn't make the tool better. Enterprise features that mostly just track how little your team uses it.
The free tier gives you 50 "agentic requests" per month, which sounds generous until you realize asking "how do I fix this Lambda timeout" burns through 3-4 requests. You'll hit the limit in a week if you actually try to use it.
The Only Three Reasons to Pay for Q Developer
You're drowning in AWS infrastructure code - If 80% of your work involves AWS API Gateway, Lambda layers, or CloudWatch alarms, Q Developer actually earns its keep. Nothing else understands AWS service configurations and limitations like this tool does.
You're trapped migrating legacy Java applications - The Java transformation features are genuinely useful and could save your team months of manual migration work. If you're facing a Java 8 to 17 upgrade across multiple codebases, this might actually pay for itself.
Your legal team runs your technology decisions - The IP indemnity coverage matters for enterprises where lawyers have more influence than engineers. Sometimes corporate risk management trumps developer experience.
The Bottom Line: Skip It Unless You're All-In on AWS
For everyone else, save your money and get Cursor for the same price, or GitHub Copilot for half the cost. Both are faster, smarter, and don't try to sell you AWS services every time you ask a question.
Amazon Q Developer isn't a bad tool - it's just an expensive, specialized tool pretending to be a general-purpose solution. If you're building on AWS and need the specific features it offers, go for it. But don't expect it to replace your existing coding workflow unless you're prepared to live entirely in Amazon's ecosystem.
