What is Amazon Nova Act
Amazon Nova Act is a general-purpose AI agent launched by Amazon AGI Labs, used to execute tasks in web browsers. Amazon Nova Act supports developers in building agent application prototypes based on the accompanying SDK, completing operations such as submitting leave requests, booking calendars, or sending automatic reply emails. Amazon Nova Act is based on decomposing complex tasks into atomic commands, combined with Playwright to directly operate the browser, improving the reliability and efficiency of task execution. Amazon Nova Act can help developers build agent applications that automate web operations, and is expected in the future to achieve more complex multi-step tasks.
Main features of Amazon Nova Act
- Decompose complex tasks: Supports developers in decomposing complex multi-step tasks into reliable atomic commands, such as searching, filling out forms, clicking buttons, etc.
- Add detailed instructions: Supports developers in adding more detailed instructions for each command, such as "do not accept insurance sales pitches" or "select a specific date," ensuring tasks are executed as expected.
- Support multi-task operations: Handles multiple tasks simultaneously, achieving parallel processing of tasks based on Python code, improving efficiency.
- Provide SDK toolkit: Amazon launched an accompanying software development kit (SDK) to help developers build agent application prototypes.
Technical principles of Amazon Nova Act
- Natural Language Processing (NLP): Based on pre-trained language models, it parses the user's intent and converts it into specific task operations. NLP technology supports multilingual processing, enabling Amazon Nova Act to provide services to users of different languages globally.
- Web automation technology: Based on automation tools (such as Selenium or Puppeteer) to simulate human web browsing behavior. Based on parsing the HTML structure of web pages, it identifies form fields to automatically fill in user-provided data. Supports extracting required information from web pages, such as restaurant reservation confirmation numbers, order status, etc.
- Machine learning and reinforcement learning: Machine learning algorithms optimize the efficiency and accuracy of task execution. Based on large amounts of training data, the model can learn the optimal task execution path. Learning technology enables Nova Act to self-optimize based on feedback during task execution. For example, if an error is encountered when booking a restaurant, strategies are adjusted based on reinforcement learning to avoid making the same mistake again.
- API calls and integration: Based on API calls, it interacts with various external services (such as restaurant reservation systems, online shopping platforms, etc.) to complete complex tasks. Amazon Nova Act provides an SDK, and developers can combine it with their own business logic to develop more customized applications.
Project address of Amazon Nova Act
- Project official website：https://labs.amazon.science/blog/nova-act
Application scenarios of Amazon Nova Act
- Office automation: Automatically submit leave requests, schedule meetings, fill out forms, and generate reports.
- E-commerce: Automatically complete online shopping, regular procurement, and price monitoring.
- Personal life management: Automatically manage schedules, healthy lifestyle arrangements, and dining reservations.
- Complex task processing: Automatically organize events (such as weddings, parties) and execute complex IT tasks.
- Enterprise business processes: Automatically handle customer relationship, supply chain management, and data analysis.