Configure your organization to use apps
Configure your organization to enable your users to create React apps.
| The new app building experience Availability | |||
|---|---|---|---|
| Cloud | Closed beta | ||
An organization admin must complete the steps in the following sections to make sure their organization is ready to use the AI app builder. The steps differ for cloud and self-hosted instances.
- Cloud
- Self-hosted
Create AI resources
Organization admins must configure one or more AI resources before builders can write prompts to create or edit apps. The following list outlines the key points of what Retool supports for AI app building:
- Admins configure their AI model provider through AI resources.
- A selection of models created by OpenAI, Anthropic, and Google Gemini are supported.
- You can also access these models through Google Vertex AI, Amazon Bedrock, and Azure OpenAI platforms.
- All organizations can use . Enterprise plan organizations can use their own keys.
- Users of OpenAI, Anthropic, and Google Gemini that also bring their own key can use an AI proxy to provide governance for their AI usage.
Custom AI provider resources are not supported in apps. Use one of the dedicated provider resources instead.
The following diagram outlines your different options for configuring AI resources.
Configure resource permissions
For users to be able to prompt AI to create apps using the AI resources you set up, they need Use access to each resource.
Add all builders to an existing or new permission group, and configure the permission group to have Use access to each relevant AI resource.
(Optional) Configure AI proxy
The following section is required only if you bring your own key and utilize an AI proxy service.
Configure a proxy
If you use a proxy to make calls to LLM providers (such as LiteLLM or LLM Proxy Server), you must configure the proxy to support models that you want to be available to your builders.
- In your proxy, configure the LLM provider that you want builders to be able to use.
- Ensure the proxy is configured to support each model that you want builders to be able to use.
- In Retool, navigate to the Resources page and open the resource corresponding to your model provider. Update the Base URL setting to point to your proxy endpoint. Repeat for any additional model providers.
Set preferred AI resource and model
Once you configure the supported AI resources that are available for building apps, you can configure your preferred resource and model. While builders have the ability to choose from any of the providers you configured, your selection will be the default for new threads.
- Navigate to Settings > AI.
- In the Apps section, configure the Preferred resource and Preferred model.
If you choose not to select a preferred resource or model, Retool automatically uses the latest Opus model.
(Optional) Configure firewall
If your organization makes use of firewall rules, you must include retool.app in its allowlist. This allows your deployment to connect to Retool's user authentication and usage reporting services.
This step is required because published apps are served at retool.app, using the following URL structure: <spaceName>--<orgName>--<app-name-slug>.retool.app. For example, an example app URL could be exampleOrg--overtime-requests.retool.app.
Create AI resources
Organization admins must configure one or more AI resources before builders can write prompts to create or edit apps. The following list outlines the key points of what Retool supports for AI app building:
- Admins configure their AI model provider through AI resources.
- A selection of models created by OpenAI, Anthropic, and Google Gemini are supported.
- You can also access these models through Google Vertex AI, Amazon Bedrock, and Azure OpenAI platforms.
- All organizations can use . Enterprise plan organizations can use their own keys.
- Users of OpenAI, Anthropic, and Google Gemini that also bring their own key can use an AI proxy to provide governance for their AI usage.
Custom AI provider resources are not supported in apps. Use one of the dedicated provider resources instead.
The following diagram outlines your different options for configuring AI resources.
Configure resource permissions
For users to be able to prompt AI to create apps using the AI resources you set up, they need Use access to each resource.
Add all builders to an existing or new permission group, and configure the permission group to have Use access to each relevant AI resource.
(Optional) Configure AI proxy
The following section is required only if you bring your own key and utilize an AI proxy service.
Configure a proxy
If you use a proxy to make calls to LLM providers (such as LiteLLM or LLM Proxy Server), you must configure the proxy to support models that you want to be available to your builders.
- In your proxy, configure the LLM provider that you want builders to be able to use.
- Ensure the proxy is configured to support each model that you want builders to be able to use.
- In Retool, navigate to the Resources page and open the resource corresponding to your model provider. Update the Base URL setting to point to your proxy endpoint. Repeat for any additional model providers.
Set preferred AI resource and model
Once you configure the supported AI resources that are available for building apps, you can configure your preferred resource and model. While builders have the ability to choose from any of the providers you configured, your selection will be the default for new threads.
- Navigate to Settings > AI.
- In the Apps section, configure the Preferred resource and Preferred model.
If you choose not to select a preferred resource or model, Retool automatically uses the latest Opus model.
Configure environment variables for Retool-managed keys
The following section is required only if you use Retool-managed keys.
To enable AI app building on self-hosted deployments using a Retool-managed key, set the following environment variables.
| Environment variable | Value |
|---|---|
LICENSE_KEY | Your self-hosted Retool license key. |
OPENAI_PROXY_HOST | Set to https://spb8yl7d3j.execute-api.us-west-2.amazonaws.com. |
OPENAI_PROXY_API_TOKEN | Contact your Retool account manager (or support@retool.com if you don't have an account manager) to generate a Retool-provided token. |
OPENAI_SECRET_KEY | Set to an empty string. |
For OPENAI_PROXY_API_TOKEN, contact your Retool account manager (or support@retool.com if you don't have an account manager) to generate a Retool-provided token. Keys provided by Retool are subject to restrictive rate limits and are not suitable for production use cases. If the rate limiting is an issue, bring your own key and set it from the Settings > AI resource page.
While this environment variable includes OPENAI in the title, you can use it for any available provider.