v3.0.157.0

🔨 Improvements

• Enhanced Element Targeting: More reliable element interaction with improved active element detection and consistent deep element location
• Better Display Compatibility: Increased screen resolution tolerance to 20% for improved compatibility across different display configurations
• Error Messaging: Clearer authentication error messages and graceful handling of unsupported viewport dimensions

🐛 Bug Fixes

• Playwright Installation: Fixed installation issues by pinning Playwright version to prevent compatibility problems
• Backend Stability: Resolved string handling and HTTP compression issues in backend operations
• Route Handling: Fix "Route already handled" error by ensuring unroute hook waits for all executing handlers to complete before removing routes, preventing race conditions during page navigation

v3.0.67.0

⚠️ Important Notice

• Support for SDK versions less than 3.0 will end on January 21, 2026

🔐 Security Updates

• SSL certificate validation is now applied to all page navigations

🚀 What's New

• Increased actuation reliability with enhanced focus detection for shadow DOM and iframe elements

🔨 Improvements

• Better authentication misconfiguration error messages with improved guidance and documentation links
• CLI error reporting now preserves full tracebacks for easier troubleshooting

🐛 Bug Fixes

• Fixed CLI deployment issues for AgentCore Runtime connectivity and duplicate workflow execution

v3.0.5.0

1. Update pyproject.toml to include the CLI templates in the generated wheel
2. Small nit to remove unused # type: ignore on boto3 client

v3.0.0.0

NovaAct 3.0.0.0

🎉 Nova Act SDK v3 is here!

This major release introduces integration with the Nova Act AWS Service, the Nova Act CLI, human-in-the loop, and (preview) tool use beyond the browser. API Keys, which provide access to the Nova Act Free Version via nova.amazon.com/act, will continue to work as well.

🚀 What's New

• Nova Act AWS Service Integration: Deploy and manage production workflows at scale using the Workflow context manager and decorator with the AWS Nova Act service
• Nova Act CLI: Rapidly deploy workflows to AWS via Amazon Nova Act and Amazon Bedrock AgentCore Runtime with simple commands - create, deploy, run, and monitor workflows directly from your terminal
• Human-in-the-Loop (HITL): Implement human supervision in your workflows with approval patterns for decision-making and UI takeover for real-time intervention via HumanInputCallbacksBase (not provided as a managed AWS service)
• Custom Tool Integration (Preview): Extend workflows beyond browser actions by integrating external tools like API calls or database queries using the @tool decorator
• Native File Upload: Upload files directly with Nova Act
• Agent Execution Time Tracking: Monitor approximate execution time to understand and optimize workflow performance

🔨 Improvements

• Enhanced Action Viewer HTML: Action Viewer now features collapsible sections and improved layout for easier debugging and workflow analysis
• Better Console Output: Added emojis and progress indicators for clearer workflow status
• Improved Actuation Reliability: Enhanced click, scroll, and type actions with better element handling and input clearing for more consistent workflow execution
• Enhanced Navigation Safety: State Guardrails now validate starting pages and navigation to prevent workflows from going off-track

⚠️ Breaking Changes

• act() API Change: In order to receive a result from act(), you now need to either pass a schema into the act() function or use the new act_get() function. If you use act() without passing a schema, response attributes a no longer returned
• SecurityOptions Update: File access control now uses allowed_file_open_paths for scoped permissions, replacing the allow_file_urls flag

Change Log

Added

• Integration with Nova Act AWS Service via the Workflow context manager and decorator
• Custom tool integration via @tool decorator to extend workflows beyond browser actions (preview)
• Human-in-the-Loop (HITL) support with approval and UI takeover patterns via HumanInputCallbacksBase
• The act cli to enable rapid deployment of Nova Act workflows to AWS AgentCore Runtime
• Time worked tracking utility to estimate and understand agent execution time
• Improved console output with emojis and additional feedback
• Native file upload support

Changed

• BREAKING: Remove response attributes from act() when called without a schema; use act_get() for structured data extraction
• BREAKING: Replace the allow_file_urls Security Option with allowed_file_open_paths to enable scoped access to files
• Redesign Action Viewer HTML with collapsible sections, expand/collapse controls, and improved visual hierarchy
• Use Delete key to clear text inputs before typing to improve reliability
• Expand State Guardrail checks to validate starting page and navigation
• Use robocopy instead of rsync when copying user data directories on Windows
• Set maximum playwright version to 1.56.0
• Improve error messages

Fixed

• Fix scrolling bugs caused by changes in Strands tool API

Security

• Improve input and output sanitization

Authors

• Aaron Challenner <aaronlmc@amazon.com>
• Andrew Werchniak <wercha@amazon.com>
• Animesh Pandey <animep@amazon.com>
• Christopher Whitten <whittech@amazon.com>
• Emile Baizel <baizele@amazon.com>
• Eric Zhou <erizhou@amazon.com>
• Matt Potok <mppotok@amazon.com>
• Sushil Bajracharya <skbajra@amazon.com>
• Tara Raj <tararaj@amazon.com>
• Ulas Kirazci <ulaskira@amazon.com>

v2.3.18.0

🔐 Security Updates

• Add configurable Security Options to block navigation to potentially unsafe URL schemes and unintended file uploads
• Tighten Content Security Policy (CSP) and sanitization in Action Viewer HTML to mitigate security risks

🚀 What's New

• Add built-in method for data extraction to increase reliability of structured extract prompts

🔨 Improvements

• Enhance scrolling accuracy for more reliable page navigation
• Improve error handling with better error classification and messaging
• Improve type definitions for better development experience

🐛 Bug Fixes

• Fix dependency version compatibility issues
• Prevent infinite loops when traversing complex DOM structures

v2.1.319.0

Release Notes

• Add state guardrail to control which URLs Nova Act can visit for enhanced security
• Improve support for select and input elements to enable more reliable interaction with dropdown menus and form fields
• Improve full page scrolling by detecting and recovering from failed scroll attempts
• Reorder operations in browser initialization to improve starting page transition time
• Safely terminate Chrome before copying user data to eliminate conflicts with default Chrome browser integration
• Strengthen protection against XSS risks, we recommend updating to this version for improved security

v2.1.124.0

Release Notes

• Connect to existing browser sessions with new cdp_use_existing_page and optional starting_page parameter to preserve browser context
• Resolve text input clearing issues across operating systems for consistent agent typing behavior
• Enhance scroll and click interactions within PDF documents
• Refine Action View HTML styling and layout for enhanced observability

v2.1.36.0

Features

• Integration with the Nova Act extension for enhanced development experience with automated environment setup, chat-based script generation, real-time debugging, and step-by-step testing capabilities

Fixes & Improvements

• Expand file upload coverage
• Fix scroll behavior when bounding box is the entire page before scrolling
• Fix model parameters not being passed to product server start-plan call
• Fix emoji decoding issues
• Improve scroll and click functionality on PDFs
• Fix scroll behavior when bounding box is the entire page (viewport dimensions)
• Improve setting of session and act ids
• Major error handling refactor with comprehensive improvements across multiple modules
• Minor logging improvements
• Relax Playwright dependency version constraints
• Add allow-origins for Chrome devtools frontend
• Interpret double and right clicks
• Additional minor fixes, improvements, and cleanup

Documentation Updates

• Add README note about cross-OS keyboard shortcuts with AgentCore Browser
• Improve documentation following error refactor
• Add README section for Nova Act extension

v2.0.357.0

Fixes & Improvements

• Strengthen type checking
• Minor refactor of telemetry module for reusability
• Remove legacy extension-related code
• Improve stop hook and S3Writer logging
• Factor browser profile and user data directory management out of default_chrome_browser feature to standardize user_data_dir behavior
• Improve scrollable element check logic
• Fix go_to_url_timeout
• Correct inaccurate ModelError raises
• Update agent_type() to insert the text if string length is > 10
• Allow scrollbars in Playwright headless mode
• Measure step server time
• Fix wait_for_page_to_settle bug that resulted in wait logic being bypassed
• Additional minor fixes, improvements, and cleanup

Documentation Updates

• Add README note about unencrypted password storage in browsers on some operating systems
• Update sample workflows

v2.0.177.0

Features

• Playwright Browser Actuation
  • Overhauled the browser actuation stack to use Playwright for improved accuracy, client-side latency, and customization

Fixes & Improvements

• Allow specifying user agent when cdp_endpoint_url is set
• Add observation_delay_ms argument to act() for customizable delays before observations, e.g., waiting for UI animations
• Improve Action Viewer log output
• Improve error messages
• Handle Chrome v138 breaking changes to extension loading
• Additional minor fixes and improvements

Documentation Updates

• README
  • Minor reorganization of Common Building Blocks section
  • Add section on integration with Amazon Bedrock AgentCore Browser Tool
  • Add notes about Playwright downloads