Speed up AI agent growth with the Nova Act IDE extension

At present, I’m excited to announce the Nova Act extension — a software that streamlines the trail to construct browser automation brokers with out leaving your IDE. The Nova Act extension integrates straight into IDEs like Visible Studio Code (VS Code), Kiro, and Cursor, serving to you to create web-based automation brokers utilizing pure language with the Nova Act mannequin.

Right here’s a fast have a look at the Nova Act extension in Visible Studio Code:

The Nova Act extension is constructed on high of the Amazon Nova Act SDK (preview), our browser automation brokers SDK (Software program Improvement Equipment). The Nova Act extension transforms conventional workflow growth by eliminating context switching between coding and testing environments. Now you can construct, customise, and take a look at production-grade agent scripts—all inside your IDE—utilizing options like pure language based mostly technology, atomic cell-style enhancing, and built-in browser testing. This unified expertise accelerates growth velocity for duties like type filling, QA automation, search, and sophisticated multi-step workflows.

You can begin with the Nova Act extension by describing your workflow in pure language to shortly generate an preliminary agent script. Customise it utilizing the notebook-style builder mode to combine APIs, knowledge sources, and authentication, then validate it with native testing instruments that simulate real-world circumstances, together with dwell step-by-step debugging of prolonged multi-step workflows.

Getting began with the Nova Act extension
First, I want to put in the Nova Act extension from the extension supervisor in my IDE. 

I’m utilizing Visible Studio Code, and after selecting Extensions, I enter Nova Act. Then, I choose the extension and select Set up. 

To get began, I have to get hold of an API key. To do that, I navigate to the Nova Act web page and observe the directions to get the API key. I choose Set API Key by opening the Command Palette with Cmd+Shift+P / Ctrl+Shift+P.

After I’ve entered my API key, I can attempt Builder Mode. This can be a notebook-style builder mode that breaks complicated automation scripts into modular cells, permitting me to check and debug every step individually earlier than transferring to the following.

Right here, I can use the Nova Act SDK to construct my agent. On the correct aspect, I’ve a Dwell view panel to preview my agent’s actions within the browser and an Output panel to watch execution logs, together with the mannequin’s pondering and actions.

To check the Nova Act extension, I select Run all cells. This may begin a brand new browser occasion and act based mostly on the given immediate.

I select Fullscreen to see how browser automation works.

One other helpful characteristic in Builder Mode is that I can navigate to the Output panel and choose the cell to see its logs. This helps me debug or overview logs particular to the cell I’m engaged on.

I also can choose a template to get began.

Moreover utilizing Builder Mode, I also can chat with Nova Act to create a script for me. To try this, I choose the extension and select Generate Nova Act Script. The Nova Act extension opens a chat dialog in the correct panel and robotically creates a script for me.

After I end creating the script, I can select Begin Builder Mode, and the Nova Act extension will assist me create a Python file in Builder Mode. This creates a seamless integration as a result of I can swap between chat functionality and Builder Mode.

Within the chat interface, I see three workflow modes accessible:

• Ask: Describe duties in pure language to generate automation scripts
• Edit: Refine or customise generated scripts earlier than execution
• Agent: Run, monitor, and work together with the AI agent performing the workflow

I also can add Context to supply related details about my lively paperwork, directions, issues, or further Mannequin Context Protocol (MCP) assets the agent can use, plus a screenshot of the present window. Offering this data helps the agent perceive any particular necessities for the automation activity.

The Nova Act extension additionally supplies a set of predefined templates that I can entry by getting into / within the chat. These templates are predefined automation eventualities designed to assist shortly generate scripts for widespread net duties.

I can use these templates (for instance, @novaAct /procuring [my requirements]) to get tailor-made Python scripts for my workflow. At launch, Nova Act extension supplies the next templates:

• /procuring: Automates on-line procuring duties (looking out, evaluating, buying)
• /extract: Handles knowledge extraction
• /search: Performs search and knowledge gathering
• /qa: Automates high quality assurance and testing workflows
• /formfilling: Completes kinds and knowledge entry duties

This extension transforms my agent growth workflow by positioning Nova Act extension as a full-stack agent builder software—an entire agent IDE for the whole growth lifecycle. I can prototype with pure language, customise with modular scripting, and validate with native testing—all with out leaving my IDE—guaranteeing production-grade scripts.

Issues to know
Listed below are key factors to notice:

• Supported IDEs: At launch, the Nova Act extension is out there for Visible Studio Code, Cursor, and Kiro, with further IDE help deliberate
• Open supply: The Nova Act extension is out there below the Apache 2.0 license, permitting for group contributions and customization
• Pricing: The Nova Act extension is out there at no cost.

Get began with Nova Act extension by putting in it out of your IDE’s extension market or visiting the GitHub repository for documentation and examples.

Glad automating!
— Donnie