You can now replay an AI browser automation without calling the model again! Nova Act records every action + URL + screenshot + DOM as JSON. Replay re-executes deterministically, validators flag any drift. Regression tests for AI agents. Finally ➡️ https://t.co/5DINSxOch1
We launched Amazon Nova Act MCP at #MCPDevSummit NYC last week, an MCP server that lets your AI coding assistant automate UI workflows through natural language.
Extract data, fill forms, handle bot-protected sites, no CSS selectors, just plain English.
One command to get started:
uvx amazon-nova-act-mcp --configure
https://t.co/ReI6nPlu1E
Kudos to Emile Baizel and Yilin Zhu from Amazon AGI Lab for creating the MCP server.
Try it out and let us know what you think!
#MCP #AIAgents #NovaAct #DeveloperTools
Looking forward to moderating this panel with research and product design leads from Amazon's AGI Lab and IBM on what it really takes to build and deploy AI agents. Come see the Nova AI Hackathon winners demo their projects and hang out with us at the Builder Loft in SF on April 9. Register or join the livestream! https://t.co/9wrrYbLizr
Amazon Nova AI Hackathon concludes with a winner celebration.
Experience demonstrations on April 9 at the AWS Builder Loft in San Francisco across Agentic AI, Multimodal Understanding, UI Automation, and Voice AI categories.
Research and product leads from Amazon's AGI Lab and IBM follow with a conversation about the future of agentic AI.
Register to attend in person or join the livestream. https://t.co/k2aLq5j5SS
Bestselling author and Gen AI instructor @anbarth talks to @JonKrohnLearns about her work at #Amazon’s #AGILabs and their newest product Nova Act, as well as where we will see the most success with AI agents and how #AI developers can reap those rewards.
Watch the episode here: https://t.co/n9jYBLTyZ5
🚀 Calling all developers: Build the future with Amazon Nova Act at the Amazon Nova AI Hackathon—compete for $40,000 in cash prizes plus $55,000 in AWS credits by creating AI agents that automate production UI workflows. Deadline: March 9, 2026. Ready to automate the future? https://t.co/WxQD9uw6tw
#AmazonNova #NovaAct #AIHackathon #Automation #BuildWithAI
🎙️ If you're looking for a new podcast to listen to this year, check out "Making a Mind" hosted by my colleague and cognitive scientist @drperszyk! ➡️ https://t.co/1GkIBM0gXP
Cognitive scientist at Amazon’s AGI Lab, @drperszyk from Amazon's AGI Lab explores the science of intelligence in "Making a Mind," a podcast featuring leading AI researchers.
Episodes 1 & 2 are now live —listen in to hear from two members of the AGI Lab technical staff, product lead, Kelsey Szot and engineer @jasonlaster11, to learn more about the evolution from LLMs to modern agents and why developing high quality training environments is as fundamental as the model itself: https://t.co/wBCbWCGNz3
Automate your smoke testing with Amazon Nova Act! This blog post shows you how to implement parallel test execution in CI/CD pipelines for faster feedback on code changes, complete with GitLab integration and best practices ➡️ https://t.co/rOAMpWmIHc
#NovaAct#UIAutomation #AmazonAGILabs
🤖 Amazon Nova Act is now GA! Build UI automation faster with less setup complexity: Prototype in the playground, refine in your IDE, and when ready, deploy to AWS. Give it a try in our new playground at https://t.co/Sy4cLuMfNS and let me know what you think! #NovaAct #UIAutomation
We’re introducing a big upgrade to Amazon Nova Act. This is our first model with large-scale RL for web tasks.
We achieve a frontier-class browser agent that's also the most cost-effective. Plus, seeing your work at a keynote is an insane feeling :)
More on the RL behind it🧵
Excited to give #AWS customers access to Claude Sonnet 4.5. This latest model from @AnthropicAI sets new standards for performance, especially for coding and building AI agents. Available today in Amazon Bedrock and coming soon to our @kirodotdev IDE to help developers build even faster. https://t.co/FyLZ5v4OC3
🚀 The Count Tokens API is now available in Amazon Bedrock to determine the token count for a given prompt or input being sent to a specific model ID prior to performing any inference. For more details, including supported models and use cases, visit: https://t.co/7GsH6Y4BRT