Amazon Nova Act Verified listing

Amazon Nova Act is a browser-based AI agent developed by Amazon's AGI SF Lab, designed to perform real-world tasks on the web through natural language commands. Leveraging Amazon's Nova foundation models, Nova Act enables developers to create agents that can interact with websites—clicking buttons, filling forms, navigating pages, and more—without the need for traditional APIs or manual scripting. Currently available as a research preview, Nova Act represents Amazon's foray into autonomous web automation, aiming to simplify complex workflows and enhance productivity.

Key Features

• Natural Language Web Automation: Instruct agents using plain English to perform tasks like online shopping, form submissions, and navigation within web browsers.
• Browser Interaction SDK: Provides a software development kit (SDK) that allows agents to interact with web elements such as drop-downs, date pickers, and pop-ups, enabling complex task execution.
• Headless Mode & Parallel Execution: Supports headless operation for background task execution and the ability to run multiple agents concurrently, enhancing scalability and efficiency.
• Integration with Amazon Nova Models: Built upon Amazon's Nova foundation models, offering capabilities in text, image, and video processing for comprehensive task handling.

Use Cases

• Automated Online Shopping: Agents can navigate e-commerce sites, select products, apply discounts, and complete purchases autonomously.
• Form Filling and Submission: Streamline processes like job applications, registrations, and surveys by automating form interactions.
• Task Scheduling and Management: Automate scheduling tasks such as booking appointments or setting reminders through web interfaces.

Technical Specifications

• Developer Access via nova.amazon.com: Available to developers through Amazon's Nova platform, providing tools and documentation for building and testing agents.
• Python-Based SDK: Utilizes a Python SDK for agent development, allowing for seamless integration with existing Python applications and workflows.
• Research Preview Status: Currently in a research preview phase, inviting developers to experiment and provide feedback for future enhancements.