You can now prompt the app building agent to create REST API or PostgreSQL resources directly from the app builder.

Describe the resource that you want to create, and the agent initiates the resource creation process. You'll be asked to provide credentials or more details where applicable.

Organizations on the Team plan can now use Retool-provided AI models for app building, in addition to the full catalog of supported models. These models, prefixed with Retool, have been tuned by Retool for enhanced performance and speed for app building.

If the organization has not set a preferred resource for AI app building, the app building agent will use a Retool-provided model.

The following Retool-provided models are supported:

• Retool Mini
• Retool Power

Organizations on the Free plan now use Retool-provided AI models for app building. These models, prefixed with Retool, have been tuned by Retool for enhanced performance and speed for app building. Access to models from other providers is now only available on paid plans.

Builders using these models will notice comparable quality and faster speeds when compared with other models previously available on the Free plan.

The following Retool-provided models are supported:

• Retool Mini
• Retool Power

The Retool CLI is now available in public beta. It enables you to build, manage, and publish Retool-native React apps from the command line and with the help of your chosen code development agent harness, like Claude Code or Codex.

With this tool, you can write and validate app code directly on your machine. You can use the CLI independently or pair it with a local development agent. Follow the installation instructions to get started, learn how the CLI works, or compare the CLI to the MCP.

Retool introduced finer controls on branches in the app builder. You can now rename branches from the Change History panel by hovering over a branch name and clicking the icon.

You can also copy the branch name, which enables builders using Source Control to transition between development in the app builder and development locally. Learn more about switching development environments.

Retool made signficant improvements to the efficiency and performance of the app building agent. Builders will notice faster agent responses and lower AI credit consumption, especially on long-running building sessions. These improvements can result in up to 50% fewer credits consumed.

The app builder now includes a debug console to help you inspect your app, surface issues, and resolve them with the agent. Open it from the status bar at the bottom of the app builder.

The console has three tabs:

• Errors: Function, build, and frontend errors. Hover a row to add its context to the chat or ask the agent to fix it, or click Fix all to resolve every error at once.
• Logs: Logs emitted by the app or the agent, including function, build, frontend, sandbox, agent, and git events.
• Timeline: A record of function runs with the elapsed time and the success of each, useful for diagnosing performance issues.

The agent has access to the same information available to builders in the debug console. This enables the agent to disagnose and address issues more quickly, and it can write more performant functions.

Enterprise organizations can use access policies to define which data each group of users can read or write on a resource. Retool enforces access policies on every query that runs against the resource. A policy can grant access to whole tables, specific columns, or specific rows.

For example, you can use an access policy to restrict an EMEA group's access on a table to only the customer rows for their own region. Retool applies that restriction to every query the group runs against the resource, so it holds regardless of which app, query, or agent the request comes from.

One policy covers queries from apps, the query library, raw SQL, and agents. Workflow queries against an enforced resource are blocked, so account for any workflows using the resource before you turn enforcement on. Retool supports access policies on PostgreSQL resources.

Access policies are enabled by default on cloud instances. On self-hosted instances, navigate to Settings > Beta and toggle on the feature flag for Enforce data security using access policies. To create a policy:

1. Open a PostgreSQL resource and select the Access Enforcement tab.
2. Turn on Data access enforcement.
3. Create a policy, choose the groups and environments it covers, and define its rules.
4. Activate the policy, then click Save changes.

Turning on data access enforcement closes the resource by default in every resource environment. Access policies apply only to the resource environments you select, so grant broad access first and narrow it once your policies are in place. Refer to Configure access policies for more details.

Access policies operate by granting access, not denying access. Refer to Access policies and Data access enforcement for more information.

You can now upload files to be used as assets in your app through the Chat tab in the app builder. Previously, uploaded files could only be used as context for the agent.

Common use cases for uploaded files as assets include:

• Embedding PDFs.
• Adding a company logo to the app header.
• Presenting images within the app.

You can now trigger workflows from functions in the new app builder. Converting classic apps that trigger workflows is now supported as a result.

To trigger a workflow from your app, open the Chat tab and prompt your agent with the workflow's name.

The triggered workflow is run on behalf of the app's authenticated user.