Reference Material reference-material

IMPORTANT
Adobe LLM Apps is currently in Beta.
Features, workflows, and UI shown here do not necessarily represent the final state of the product. To join the Beta, send an email to llm-apps-beta@adobe.com.

This section provides field-level reference for action configuration in the Adobe LLM Apps UI.

Action parameters

Input parameters are the values the LLM platform (ChatGPT, Claude) sends to your action handler. The model extracts them from the user’s message and maps them to these fields automatically.

Property
Description
Name
The parameter identifier (for example, category, query)
Type
String, Number, Integer, or Boolean
Description
A human-readable explanation — the LLM platform uses this to extract the right value
Required
If checked, the model must provide this parameter before invoking the action

File parameters

File parameters carry file objects with download_url and file_id properties. Define input field names that should receive file data when a user uploads a file in the conversation.

Metadata fields

Basic information

Field
Required
Description
Action name
Yes
Identifier for the action (for example, Search Products)
Description
Yes
Explanation of what the action does — the LLM platform uses this to decide when to invoke it

Annotations

Optional hints that describe the action’s behavior:

Annotation
Description
Destructive hint
The action modifies or deletes data
Idempotent
Calling the action multiple times with the same arguments has the same result
Open world hint
The action interacts with external systems
Read only hint
The action only reads data, never writes

OpenAI metadata

Field
Max length
Description
Invoking status text
64 chars
Message shown in the LLM platform while the action runs (for example, Loading products … )
Invoked status text
64 chars
Message shown after the action completes (for example, Products loaded … )

Visibility

Toggle
Description
Expose to AI model
The action can be invoked by the AI model during conversations
Show as widget in app surface
The action renders a visual widget in the app

Widget information

Field
Description
Type
Widget technology — currently EDS
Widget domain (sandbox origin)
Origin where the widget is hosted; must be unique per app
Prefers border
If checked, the widget renders inside a bordered card in the LLM platform

Template URLs

Field
Description
Script URL
Entry point script — https://main--<repo>--<owner>.aem.live/scripts/aem-embed.js. Shared across all actions
Widget embed URL
EDS page for this action — https://main--<repo>--<owner>.aem.live/eds-widgets/<action-name>. Unique per action

CSP configuration

Content Security Policy controls which external domains the widget iframe may contact. Every external domain must be explicitly allowlisted.

Directive
Description
Resource domains
Domains for static assets — images, fonts, scripts, styles
Connect domains
Domains the widget may contact via fetch, XHR, or WebSocket
Frame domains
Origins allowed for nested iframes; triggers stricter app review
Redirect domains
Trusted targets for openExternal redirect links (ChatGPT-specific)
Base URI domains
The base-uri CSP directive (MCP Apps SDK only, not ChatGPT)

Permissions

Hardware and browser APIs the widget is allowed to access. These map to the iframe permission policy.

Permission
Description
Camera
Access the device camera
Microphone
Access the device microphone
Geolocation
Access the user’s location
Clipboard
Read from or write to the clipboard
recommendation-more-help
llm-apps-help-main-toc