Amazon Q Developer Review: AWS's AI Coding Bet Pays Off - With Caveats

Every major cloud provider now has an AI coding assistant, but Amazon Q Developer is the one that most clearly reveals its parent company's strategy. This isn't a general-purpose code completion tool dressed up for marketing. It's an AWS recruitment engine that also happens to write decent code. That framing matters, because understanding what Amazon Q Developer is optimized for - and what it's not - is the key to deciding whether it belongs in your workflow.

After several weeks of testing across VS Code, JetBrains, and the CLI, I came away impressed by its AWS-specific capabilities and surprised by how much the free tier offers. But I also hit real limitations that keep it from competing with the best all-rounders.

From CodeWhisperer to Q Developer

A quick history lesson. Amazon Q Developer is the rebranded and significantly expanded successor to AWS CodeWhisperer, which launched in 2022. The April 2024 rebrand wasn't just cosmetic - it brought agent capabilities, a CLI experience, AWS Console integration, and the /transform feature for automated Java upgrades. The product has matured considerably since then, with new agents for /doc, /test, and /review arriving in late 2024 and an enhanced agentic CLI powered by Claude 3.7 Sonnet rolling out more recently.

The naming is confusing, though. "Amazon Q" is also the name of AWS's business intelligence assistant (Amazon Q Business), and the two products share a brand but serve completely different audiences. If you are a developer looking for a coding assistant, make sure you are reading about Amazon Q Developer specifically.

The IDE Experience

Amazon Q Developer is available as extensions for VS Code, the full JetBrains family (IntelliJ, PyCharm, GoLand, WebStorm, and others), and Eclipse in preview. It supports over 25 programming languages including Python, Java, JavaScript, TypeScript, Go, Rust, C#, and Kotlin.

In VS Code, the extension installs cleanly and authenticates through either AWS Builder ID (free) or IAM Identity Center (for Pro tier enterprise setups). The initial experience feels similar to other AI assistants: you get inline code suggestions as you type, a chat sidebar for questions, and slash commands for specialized tasks.

Amazon Q Developer integrates into familiar IDE environments with inline suggestions, chat, and specialized agent commands.

Code completions are decent but not exceptional. For straightforward patterns - boilerplate, control flow, common library usage - the suggestions are reliable. Amazon claims industry-leading acceptance rates, with BT Group reporting 37% and National Australia Bank hitting 50%, the latter climbing to 60% with codebase customization. Those are strong enterprise numbers, but in my hands-on testing, general-purpose completion quality sits a tier below what Cursor or GitHub Copilot deliver. Completions often feel more conservative - shorter, less willing to produce full function bodies, and occasionally slow to appear.

Where Amazon Q Developer starts differentiating is in AWS-specific contexts. Ask it to create a CloudFormation template, write a Lambda handler with proper error handling and logging, or create an IAM policy with least-privilege principles, and the quality jumps noticeably. The model has clearly been trained on AWS documentation and best practices. It knows which S3 bucket policies are dangerously permissive, it creates CDK constructs with sensible defaults, and it understands DynamoDB access patterns in a way that a generic coding model does not.

Agent Mode - The Headline Feature

The /dev agent is where Amazon Q Developer shows the most ambition. You describe a feature in natural language, and the agent breaks the task into steps, builds code across multiple files, generates tests, and presents its work for review. This is similar to agentic coding in Copilot and Cursor, and the results are mixed in much the same ways.

For well-scoped tasks - "add input validation to this API endpoint" or "create a DynamoDB table with a GSI for querying by email" - the agent produces usable output that needs minor tweaking. For more complex requests, it often delivers incomplete implementations or creates code that conflicts with existing patterns in your project. The context window limitation is real: the agent mostly works with the files you have open, so it can miss project-wide conventions and dependencies.

The specialized agents add real value:

• /review scans your code for security vulnerabilities, anti-patterns, and logical errors using a combination of generative AI and rule-based detectors built on years of Amazon's internal security practices. It catches exposed credentials, SQL injection risks, and infrastructure misconfigurations that static analysis tools might miss.
• /transform automates Java version upgrades - from Java 8 or 11 to Java 17, for example. Amazon claims a 85% success rate on large open-source projects with over 100,000 lines of code, and customers have reported upgrades that would take weeks completed in minutes. This is a genuinely useful capability for enterprises drowning in legacy Java code.
• /doc creates documentation including data flow diagrams, and /test creates unit tests iteratively for low-coverage codebases.

The free tier gives you 50 agentic requests per month. That sounds generous until you realize that complex tasks often require multiple back-and-forth interactions. Power users will hit the limit within a week or two.

The CLI - An Underrated Feature

The Amazon Q Developer CLI deserves more attention than it gets. It offers three capabilities: command autocompletion with a graphical dropdown menu that works across hundreds of popular CLI tools (git, npm, docker, aws), inline ghost-text suggestions for shell commands, and an interactive chat powered by Claude 3.7 Sonnet with step-by-step reasoning.

The q translate command converts natural language to shell commands, which is useful for those moments when you know what you want to do but cannot remember the exact aws CLI syntax for describing a CloudWatch alarm with specific dimensions. The chat can access tools on your system including compilers, package managers, and the AWS CLI itself, making it a truly capable terminal companion.

For developers who live in the terminal, this is arguably Amazon Q Developer's most practical feature. It isn't flashy, but it reduces friction in real workflows.

Amazon Q Developer's deep integration with AWS services makes it most valuable for teams building on Amazon's cloud infrastructure.

Security Scanning - Baked In, Not Bolted On

Security is where Amazon Q Developer truly stands out from most competitors. The integrated scanning covers three areas: static application security testing (SAST) for vulnerabilities like resource leaks and cross-site scripting, secrets detection for hardcoded passwords and connection strings, and infrastructure-as-code scanning for misconfigured AWS resources.

The auto-scan feature runs in the background as you write, flagging issues in real time rather than waiting for a CI pipeline. Some detections come with one-click fix suggestions that apply the remediation directly in your IDE. The detector library draws from Amazon's internal security practices and the CodeGuru detection catalogue.

For teams that need security scanning integrated into the developer workflow rather than gated behind a separate tool, this is a meaningful differentiator. Most competitors either lack built-in security scanning or offer it as an add-on. Amazon Q Developer includes it in both the free and Pro tiers.

Customization and Enterprise Features

Pro tier users ($19/month per seat) can customize Amazon Q Developer's suggestions based on their private codebase. You connect a code repository through AWS CodeConnections or point it at a S3 bucket containing your source code, and the model learns your organization's patterns, naming conventions, and internal libraries. National Australia Bank reported their acceptance rates climbed from 50% to 60% after enabling customization.

The customization currently works for Java, JavaScript, TypeScript, and Python, and requires at least 2 MB of source code. Your code stays private - AWS does not use it to train the foundation model, and the inference endpoint is isolated to your organization.

Enterprise governance is solid. The Pro tier adds an admin dashboard for user and policy management through IAM Identity Center, usage tracking, and automatic opt-out of data collection. IP indemnity is included, which matters to legal teams. For organizations that need to demonstrate compliance, these controls are more mature than what most coding assistants offer.

Pricing - The Free Tier Is Genuinely Generous

The free tier is remarkably complete for individual developers. Unlimited code suggestions, 50 agentic requests, and built-in security scanning - no credit card required. Compared to GitHub Copilot's $10/month individual plan or Cursor's $20/month Pro tier, getting started with Amazon Q Developer costs nothing.

The Pro tier at $19/month is competitive with Copilot Business ($19/month) and undercuts Cursor Business ($40/month), though direct price comparisons miss the point. You aren't paying for equivalent products - you are paying for different strengths.

What Holds It Back

The honest assessment: Amazon Q Developer is a strong number-three in a market with clear leaders.

General code quality trails the competition. In side-by-side testing, both Cursor and GitHub Copilot create higher-quality completions for general-purpose coding. Amazon Q's suggestions are more conservative, more likely to be line-level rather than function-level, and occasionally slow to appear. A recent Visual Studio Magazine comparison testing agentic capabilities in VS Code found Copilot delivering higher adoption and acceptance rates.

Context limitations are real. The conversational memory in chat is limited primarily to the active file. Close a file and the assistant forgets the functions and patterns defined there. For large projects with deep dependency chains, this means you spend time re-explaining context that a tool like Cursor would retain through its full project indexing.

The AWS gravity pull is strong. If you aren't building on AWS, much of Amazon Q Developer's value proposition evaporates. Its AWS-specific knowledge is its biggest differentiator, but it's also its biggest limitation. Multi-cloud teams or those building on GCP or Azure will find less value here. As one Gartner reviewer noted, poor integration with non-AWS projects limits the tool's effectiveness.

Agent mode is not finished. Complex agentic tasks frequently produce incomplete results, and the agent has a tendency to overwrite existing patterns rather than work within them. The 50-request monthly limit on the free tier compounds this - you burn through requests quickly when debugging the agent's own output.

Who Should Use Amazon Q Developer

The ideal user is an experienced developer on a team deeply invested in AWS. If your daily work involves Lambda functions, DynamoDB tables, CloudFormation templates, CDK stacks, and IAM policies, Amazon Q Developer offers context-specific assistance that no competitor matches. The security scanning adds genuine value, the free tier lets you assess without commitment, and the Pro tier's customization and governance features address real enterprise needs.

If you're looking for the best general-purpose AI coding assistant, look elsewhere. Cursor remains the leader for deep project understanding and multi-file editing, and GitHub Copilot offers a more polished experience across varied codebases. For CLI-focused workflows, Claude Code provides a more capable terminal-native experience.

For teams considering their options, our best AI coding assistants roundup covers the full landscape. And if you're new to the space entirely, our getting started guide will help you navigate the choices.

Verdict: 7.4/10

Amazon Q Developer is a capable AI coding assistant that excels in its niche. The AWS integration is genuine, the security scanning is best-in-class among coding assistants, the /transform agent solves a real enterprise pain point, and the free tier is one of the most generous in the market. But general-purpose code completion and chat quality lag behind the leaders, context management needs work, and the product's value drops sharply once you step outside the AWS ecosystem. For AWS-native teams, it's a strong addition to the toolkit. For everyone else, it's a solid free option that doesn't quite justify switching from a better all-rounder.

Sources

• Amazon Q Developer - Official Features Page
• Amazon Q Developer Pricing
• Amazon Q Developer Review - InfoWorld
• Comparing Amazon Q and GitHub Copilot Agentic AI in VS Code - Visual Studio Magazine
• BT Group Case Study - Computer Weekly
• GitHub Copilot vs Amazon Q Enterprise Bakeoff - Faros AI
• Amazon Q Developer Agent Capabilities - AWS Blog
• Enhanced CLI Agent in Amazon Q Developer - AWS DevOps Blog
• Code Security Scanning with Amazon Q Developer - AWS DevOps Blog