Starting a cloud project usually involves several steps before developers can deploy their first application. They need an AWS account, must configure authentication, install the AWS CLI, create or select an IAM identity, and provide the application with the permissions required to access cloud resources.
For developers using AI coding agents, this setup can interrupt the development workflow. An agent may be ready to generate an application and deploy it, but it cannot access AWS until the developer completes account creation and credential configuration.
AWS is addressing this problem with a new account creation experience designed to work with AI coding tools.
Through the AWS Agent Toolkit and the AWS CLI, a coding agent can initiate the AWS sign-up process directly from its development environment. The browser opens for the developer to complete the required registration and authentication steps. Afterward, AWS provides temporary credentials that the agent can use through supported AWS tools.
This removes the need to manually generate and share long-lived access keys with the coding agent.
The important distinction is that AWS is not removing authentication. It is connecting account creation, browser-based authentication, and temporary credential provisioning into a more integrated development workflow.
This article explains how the experience works, how to get started, and what developers should consider before allowing an AI agent to create and manage AWS resources.
Why AWS Account Setup Can Interrupt AI Development
Consider a developer who wants to build a serverless application using an AI coding agent.
The developer asks the agent to create an API, deploy a Lambda function, configure storage, and make the application accessible through an endpoint.
The agent can generate the application code, but deploying it requires access to an AWS account.
In a traditional workflow, the developer must:
Visit the AWS website and create an account.
Complete identity and account verification.
Configure authentication for programmatic access.
Install and configure the AWS CLI.
Create or select an IAM identity.
Provide the development environment with appropriate credentials.
Verify that the credentials can access the required AWS services.
For developers who already work with AWS, these steps are familiar. For someone starting a new project, they can become an unnecessary barrier before the actual development work begins.
The problem becomes more noticeable when using AI coding tools such as Claude Code, Codex, Cursor, or other supported agent environments.
The agent can generate code and execute terminal commands, but it still needs an authenticated AWS identity to interact with cloud services.
AWS's new account creation workflow connects these steps so that the developer can start the sign-up process from the coding environment instead of manually switching between several tools.
Understanding the New AWS Account Creation Experience
AWS announced agentic account creation for coding harnesses in November 2026.
The feature allows supported AI coding environments to initiate the new AWS sign-up experience when a developer does not already have an AWS account.
The basic workflow looks like this:
Developer starts an AWS project
|
v
AI coding agent initiates AWS sign-up
|
v
Browser opens AWS registration
|
v
Developer completes registration
|
v
AWS provisions the first project
|
v
Temporary credentials become available
|
v
Agent connects to AWS tools
|
v
Agent can build and deploy resources
The developer still participates in the registration and authentication process.
The agent does not independently establish the developer's identity or bypass AWS account verification.
Instead, it initiates the workflow and uses the authenticated session made available after the browser-based process is completed.
This is an important security distinction. The goal is to remove unnecessary credential handling, not to remove the developer's control over account creation.
Understanding AWS Projects and the New Sign-Up Model
The new AWS experience introduces a project-based way to organize cloud resources.
A project contains an AWS account and settings for sharing resources with collaborators. Developers can manage projects, invite team members, and configure other settings through AWS Settings.
When signing up through this experience, AWS also creates an AWS Builder ID account, which represents the developer as an individual.
Supported sign-in options include existing Google, GitHub, Apple, and Amazon identities.
After completing registration, AWS provisions the first project in an available AWS Region.
This is different from the traditional advanced AWS sign-up experience, where developers have more direct control over account configuration and can access advanced capabilities immediately.
The new experience is intended to make getting started easier through preconfigured defaults.
However, developers working with regulated workloads, requiring complete control over where content is stored, or needing advanced networking and security services immediately should evaluate the advanced sign-up option instead.
The choice matters because not every AWS service or account configuration is necessarily available through the same project experience.
Starting AWS Account Creation From an AI Coding Agent
The AWS Agent Toolkit provides the connection between supported coding agents and AWS.
It brings together AWS documentation access, agent skills, and AWS account connectivity.
When a developer does not have an AWS account, the toolkit can guide the agent through the account creation process.
The workflow begins by configuring the AWS Agent Toolkit in the coding environment.
AWS provides an official setup guide and plugin-based integration for supported agent environments.
After the toolkit is installed, the developer can ask the agent to connect to AWS.
For example, a developer might provide this instruction:
I want to create a new serverless application on AWS.
Check whether an AWS project is already configured.
If I do not have an AWS account, help me start the AWS sign-up
process using the official AWS workflow.
Do not ask me to provide access keys or secret keys.
After authentication, verify the AWS account and Region
before creating any resources.
If the agent detects that AWS authentication is missing, it can initiate the appropriate sign-up or login workflow.
The browser then opens the AWS registration page, where the developer completes the required steps.
Once registration and project selection are complete, the AWS CLI receives credentials that can be used by the agent.
The developer can then continue the original development task.
Using the AWS CLI to Create an Account
Developers can also start the new sign-up experience directly from the terminal.
The AWS CLI provides the aws signup command for this purpose.
First, install or update the AWS CLI.
The aws signup command requires AWS CLI version 2.41.0 or later.
Check the installed version:
aws --version
If the required version is installed, start registration:
aws signup
The AWS CLI opens the default browser and directs the developer to the AWS sign-up experience.
The terminal command remains active while the developer completes registration.
After signing up, the developer selects the newly created project in the AWS browser flow.
When the process completes, the CLI receives credentials for that project.
To use a named profile instead of the default profile, provide the profile option:
aws signup --profile MyFirstProject
The profile provides a separate configuration entry for the project.
For developers using an earlier AWS CLI version, AWS documents aws login as an alternative entry point. With supported versions, this command opens the sign-in experience, where the developer can choose the option to sign up.
aws login --profile MyFirstProject
The distinction is straightforward:
Both commands use a browser-based authentication process rather than requiring developers to manually create long-lived access keys.
How the Agent Gets AWS Credentials Without Long Lived Keys
One of the most important parts of this experience is credential management.
Traditional AWS CLI configuration often involves storing an access key ID and secret access key.
For example, developers might previously have used:
aws configure
The CLI then requested the access key ID, secret access key, default Region, and output format.
This approach can create problems when credentials are copied into project files, environment variables, scripts, or agent configuration.
The new workflow uses temporary credentials obtained through browser-based authentication.
After successful sign-up or login, the AWS CLI stores the required authentication information in its profile and manages credential renewal.
The agent can use the configured profile through supported AWS CLI, AWS SDK, and AWS Tools for PowerShell workflows.
For example:
aws sts get-caller-identity \
--profile MyFirstProject
This command returns the identity associated with the active credentials.
The developer does not need to paste an access key or secret key into the agent's conversation.
It is important to understand that the credentials are not necessarily absent from the local machine. The AWS CLI maintains the authentication information required to use and renew the session.
The security improvement is that developers do not need to create and distribute permanent IAM access keys for this workflow.
Understanding Credential Expiration and Renewal
Temporary credentials are not permanent.
For the new AWS project connection experience, AWS documents a 12-hour access period.
Developers can renew access using:
aws login --profile MyFirstProject
The browser opens automatically, and the developer selects the project to continue the authentication process.
AWS can renew credentials for up to 90 days before requiring the developer to grant access again through a browser.
This provides a balance between convenience and session control.
For example, a developer can connect an AI coding agent in the morning, work on an application, and allow the agent to use the project's credentials during that session.
When the credentials expire, the developer can renew the session rather than generating another permanent access key.
One important point is that credential expiration does not automatically mean that every resource created during the session is removed. AWS resources remain in the account until they are deleted or otherwise changed.
The credentials control access to resources, not the lifetime of the resources themselves.
Connecting Multiple AWS Projects to Different Coding Workflows
Developers often work on more than one application.
For example, a developer might maintain separate AWS projects for:
A personal serverless application.
A development environment for a client.
An internal application used for testing.
The AWS CLI supports named profiles so that each project can have a separate configuration.
For example:
aws login --profile PersonalProject
aws login --profile ClientDevelopment
aws login --profile InternalTesting
Each profile represents a separate project connection.
When authenticating, the developer must select the correct project in the AWS session selection experience.
This is important because selecting the wrong project can cause the agent to create resources in an unintended AWS environment.
Before executing deployment commands, verify the active identity:
aws sts get-caller-identity \
--profile ClientDevelopment
The result should identify the expected AWS account and caller identity.
Also verify the configured Region:
aws configure get region \
--profile ClientDevelopment
A profile name is only a local label. It does not prove which AWS project is currently selected.
Always verify the actual account and Region before allowing an agent to make changes.
What the AWS Agent Toolkit Adds to the Workflow
Account creation and authentication are only part of the integration.
The AWS Agent Toolkit also provides capabilities that help coding agents work with AWS more effectively.
It combines three important components.
AWS Documentation and API References
An AI model may not always have the latest information about AWS services, APIs, or configuration changes.
The AWS documentation integration allows the agent to retrieve current documentation rather than relying exclusively on information learned during model training.
This is useful when generating infrastructure code, configuring services, or troubleshooting an AWS deployment.
AWS MCP Server
The Model Context Protocol (MCP) provides a way for AI agents to access external tools and services.
The AWS MCP Server gives supported agents access to AWS-related tools.
It can help agents retrieve service information and perform AWS operations using the authenticated identity.
The permissions available to the agent depend on the AWS identity and project it uses.
AWS Agent Skills
The toolkit also includes task-specific skills that guide the agent through common AWS workflows.
Examples include serverless application development, infrastructure as code, and IAM configuration.
These skills provide structured instructions that help the agent follow established AWS practices.
The combination of documentation, tools, and skills makes the agent more capable than a workflow based only on terminal commands.
However, these capabilities do not remove the need for developers to review infrastructure changes and verify the permissions being used.
Security and Cost Considerations
Allowing an AI agent to create and manage cloud resources introduces responsibilities that developers should not overlook.
Verify the AWS project before deployment
Always confirm the account ID, caller identity, and Region before allowing the agent to create resources.
Use:
aws sts get-caller-identity
Do not rely only on the profile name or the agent's statement about the selected environment.
Review resource permissions
Temporary credentials do not automatically mean least-privilege access.
The agent can perform operations allowed by the permissions associated with its project and identity.
Review the available services and permissions before using the agent for infrastructure management.
Understand resource costs
The developer remains responsible for resources created by the AI coding agent.
An agent might create Lambda functions, storage resources, databases, or other services that incur charges.
The new AWS experience provides plan and spend-limit options where applicable. Developers should review their plan, pricing, and available spending controls before deploying resources.
Do not expose authentication information
Avoid printing credential files, session tokens, or authentication details in terminal output.
The agent should use the configured AWS CLI or SDK credential provider rather than receiving secrets through prompts or source code.
Review infrastructure changes
An agent may generate an infrastructure configuration that creates more resources than the application actually needs.
Review deployment plans, resource definitions, and permissions before applying changes, particularly in shared or production environments.
When Should Developers Use the New AWS Sign-Up Experience?
The new workflow is especially useful for developers who are starting AWS projects and want to work directly from an AI coding environment.
It can reduce setup friction for:
Developers experimenting with serverless applications.
Students learning AWS through practical projects.
Developers building prototypes with AI coding tools.
Teams creating separate development projects.
Engineers who want to avoid managing permanent access keys for local development.
However, the new sign-up experience is not a universal replacement for every AWS account setup.
Organizations with existing account governance, regulated workloads, advanced networking requirements, or specific security and compliance controls should evaluate whether the new project model meets their requirements.
The traditional advanced sign-up process may be more appropriate in those situations.
Summary
AWS is making it easier for AI coding agents to move from application development to cloud deployment by integrating account creation and authentication into the agent workflow.
Through the AWS Agent Toolkit and the AWS CLI, a developer without an AWS account can start registration directly from a supported coding environment. The browser handles the required sign-up and authentication steps, and the CLI receives temporary credentials for the selected AWS project.
The developer does not need to generate permanent access keys and paste them into the agent's configuration.
The workflow also supports named profiles for multiple projects, temporary credential renewal, AWS documentation access, MCP integration, and task-specific agent skills.
However, convenience should not replace security. Developers remain responsible for the resources created by their agents, the permissions granted to those agents, and any associated AWS charges.
The most useful change is that account creation, authentication, and cloud development can now be part of one connected workflow. Developers can spend less time managing credentials and more time building applications while retaining control over the account and authentication process.
Join the conversation! Your thoughts help the community grow.