Amazon Nova Act

What it does

Amazon Nova Act is a service and developer toolset for building AI agents that automate browser-based UI workflows. You describe steps in natural language, Python code, or both, and the agent acts in a hosted or local web browser, breaking tasks into shorter actions. Outputs include executed browser actions, extracted data, and run traces with step-level observability in the AWS console.

People use it to explore automation ideas in the playground, then move to the Python SDK and IDE extension to debug locally with breakpoints and assertions. They then deploy with the CLI to Amazon Bedrock AgentCore Runtime. Documented use cases include web QA testing, form data entry, data extraction from dashboards, and end-to-end checkout flows. Run traces and observation artifacts stored in S3 support monitoring and review.

It suits developers and teams automating web tasks who are comfortable with Python. The playground needs only an Amazon account, but the CLI and AWS console require an AWS account and configured AWS credentials. Playground and API-key use follow the nova.amazon.com Terms of Use, while IAM-authenticated and deployed use follow AWS service terms. Request a quote or check AWS pricing pages for deployed-service costs.

What it can do

• Natural-language and Python-based definition of browser automation agents
• Browser-based execution of UI workflows in a hosted playground
• Local debugging with breakpoints, assertions, and iteration in the Python SDK
• Packaging and deployment of scripts as remote workflows via the CLI
• Workflow run monitoring with execution traces (runs, sessions, acts, steps) in the AWS console
• Access to observation artifacts stored in S3

Works with

• Amazon Bedrock AgentCore Runtime
• Amazon S3
• Amazon ECR
• AWS IAM

Example workflows

Form data entry across portals

Input: a list of records and target web portal URLs. Steps: define a natural-language or Python agent that navigates to each form, fills fields, and submits. Output: submitted form entries and per-step execution traces in the console.

Web QA test run

Input: a set of UI test cases described in natural language. Steps: write agent scripts for each case, run them locally in the SDK, then deploy with the CLI for scheduled or parallel runs. Output: pass/fail results with execution traces and S3-stored artifacts.

Before you choose

• Playground and API-key use is governed by the nova.amazon.com Terms of Use, while IAM-authenticated and deployed use is governed by AWS service terms.
• The CLI automates containerization, ECR, S3, and IAM role creation, so review the resources and costs created in your AWS account before deploying.
• Confirm pricing, quotas, and regional availability for the deployed service with AWS before production use.

Listed under

• Coding
• App builders
• Code review and testing
• AI agents
• Browser automation
• Developers
• Enterprise
• Form filling
• No-code

Get started

Access the playground with an Amazon account; SDK install is via pip install nova-act.

1. Sign in to the playground at nova.amazon.com/act with an Amazon account and test an automation idea.
2. Install the SDK with pip install nova-act and authenticate with an API key for quick experimentation or AWS IAM for production work.
3. Configure AWS credentials via the AWS CLI or profiles before using the CLI to deploy workflows.

Pricing was not confirmed in the reviewed information. Check AWS pricing pages before deploying workflows, since AWS resources may incur charges.

We couldn’t find published prices for Amazon Nova Act. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 28 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Understand the available interfaces for using Nova Act - Amazon Nova Act ↗docs.aws.amazon.com
• Amazon Nova Act Documentation ↗aws.amazon.com

Company website

nova.amazon.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Web search

Confirms the official site

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 28 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.