Edit the Blog Post Header component above this, then place your content here, then fill out the Related Articles section below, if possible. You should have at least one related article and ideally, all three.
Feel free to add supplementary content to the sidebar at right, but please retain the Twitter component (it can live at the bottom of your added content).
This text component can be edited or deleted as necessary.
Related articles only have an image (the squarer version of the banner) and a title. This text can be deleted. Place 1-3 related articles.
Integrating In-App Purchasing (IAP) with an app, or moving it to a newer Android API level, usually means weeks of reading documentation, wiring up SDK calls, and testing edge cases before you can ship. Amazon Devices Builder Tools (ADBT) gives your AI coding agent the Fire TV knowledge for integrations and step-by-step workflows to do that work with you, in a fraction of the time.
ADBT is a free, publicly available set of tools for Fire TV development, including an MCP server and agent skills. It works with Claude Code, Cursor, Kiro, GitHub Copilot, and other MCP-compatible agents.
You can use ADBT to get started with app development, but once you’ve done that, you can also use it for:
To install ADBT and configure it for your agent, run this command in your project folder:
npx -y @amazon-devices/amazon-devices-buildertools-mcp@latest init-context
The command installs the MCP server, steering documents, and skills for the agent you choose. After that, describe what you want in your agent's chat, and your agent picks the right workflow. For details, see the documentation on how to Set Up Amazon Devices Builder Tools for AI-Powered Development.
Choose your AI agent when you run init-context.
When setup finishes, init-context shows a summary and suggested next steps.
To start, ask your agent something like:
Help me integrate IAP in my app
Claude Code starts the IAP setup phase after checking the development environment.
ADBT walks your agent through six phases, in order. You stay in control at each step, and your agent stops to ask you when it needs a decision.
The generated code follows the recommended purchase flow, so you don't have to piece it together from the API reference. ADBT also marks the places that need your own implementation. For example, the generated receipt storage is a reference stub that you replace with your own persistence solution, and receipt verification is stubbed until you connect it to your RVS server.
The generated IAP files. The receipt storage is a reference stub, marked with a TODO to replace it with your own persistence.
If your Fire OS app still uses the legacy IAP v2.0 library, or an older version of the Amazon Appstore SDK 3.x, ADBT can migrate it to the latest Appstore SDK 3.x. Support for IAP v2.0 is being discontinued, so new submissions should use the Appstore SDK.
To start, ask your agent something like:
Migrate my Fire OS app to the latest Amazon Appstore SDK
Your agent works through the migration in order, and stops to ask you before any change with a wide reach:
Your purchase, receipt, and fulfillment code stays the same. If your project doesn't include its public key yet, download AppstoreAuthenticationKey.pem from the Developer Console and add it to app/src/main/assets/. Without it, purchases work in sandbox testing but fail in production. For the steps, see the documentation on how to Integrate the Appstore SDK.
After the migration, your agent can continue with on-device testing, unit tests, and submission, the same phases it uses for a new IAP integration. To learn more about the change, see the documentation on Transitioning from IAP v2.0 to the Appstore SDK.
Claude Code moves a Fire OS app from Appstore SDK 3.0.2 to the latest 3.x, and confirms that the app still builds.
Raising your app's targetSdkVersion opts it into the behaviors of a newer Android API level, which can change how it runs, builds, or installs. The Fire TV Android SDK upgrade workflow helps you move a Fire OS app across API levels 30 through 36 safely.
To start, ask your agent something like:
Upgrade my Fire TV app from targetSdkVersion 30 to 35
Instead of jumping straight to the target level, ADBT moves one API level at a time, and handles one behavioral change at a time within each level:
Before changing anything, Claude Code proposes an upgrade path from API 30 to 35, one level at a time, and asks for your approval.
Claude Code tests the build-tooling changes in a temporary copy of the project, then shows the exact diff and waits for approval.
Every edit needs your explicit approval, and a prompt or tool permission never counts as approval. Each change also traces back to the documentation for the API level that introduced it, so you can review and repeat the upgrade later.
The SDK & CLI workflow lets your agent handle the Vega SDK and command-line tools for you, so you can stay in your editor. Ask your agent to:
Install/update SDK from AI agents
Build my app
Run my app on virtual device
Run my app on physical device
Run KPI Visualizer and get performance results
Your agent checks prerequisites, installs or updates the Vega SDK, runs the right CLI commands, and reports the results back in the chat. That includes performance results from the KPI Visualizer, so you can measure your app against Fire TV performance targets as you build.
Claude Code runs KPI Visualizer cold-start tests on a Fire TV device and compares Debug and Release builds against the Fire TV targets.
With ADBT, your AI coding agent can integrate IAP into your Vega OS or Fire OS app, move your Fire OS app to a newer Android API level, and build, run, and measure your Vega app, all from your agent's chat. You don't need to wait for help from an Amazon engineer: ADBT is free, self-serve, and available today.
To get started, run npx -y @amazon-devices/amazon-devices-buildertools-mcp@latest init-context in your project folder, and then ask your agent for help. To see some workflow and example prompts, see the Amazon Devices Builder Tools ReadMe. Have feedback? Tell your agent "I want to provide feedback" or share it in the Amazon Developer Community.