ChatGPT Apps SDK

The ChatGPT Apps SDK is the developer framework OpenAI announced at DevDay 2025 for building interactive applications that run inside ChatGPT. Unlike the earlier plugin system (deprecated in 2024), Apps can render rich UI - cards, carousels, forms, even full-screen experiences - and persist context across a conversation.

The SDK is built on the Model Context Protocol (MCP), the open standard Anthropic originated and OpenAI adopted. This means an MCP server written for Claude or another MCP-compatible host can be adapted into a ChatGPT App with relatively modest additional work, and the same server can serve multiple surfaces.

Launch partners at DevDay 2025 included Canva, Expedia, Spotify, Figma, Booking.com, and Zillow. The stated design goal: users should not have to leave ChatGPT to book travel, create a design, browse listings, or buy something. The app renders inline inside the conversation, inherits the conversational context, and can take action on the user's behalf.

Apps complement rather than replace the Agentic Commerce Protocol (ACP). ACP handles the transaction layer - the open standard for enabling Instant Checkout. Apps handle the experience layer - how a merchant's catalog is browsed and explored before the transaction. A retailer with both a ChatGPT App and ACP integration controls both the browse and the buy inside ChatGPT.

Before Apps, a retailer's presence inside ChatGPT was limited to whatever ChatGPT's retrieval layer chose to surface - text mentions, product cards via the Shopping feed, or citations from the broader web. Retailers could optimize the inputs (ACO, GEO, schema) but could not own a branded surface inside the conversation.

The Apps SDK changes that. A retailer with a ChatGPT App gets:

• An owned, branded surface inside ChatGPT - the brand name, visual identity, and navigation live inside the conversation, not subordinate to it
• Interactive UI for product discovery - carousels, filters, comparison tools that work inside the chat flow
• Conversational context - the app inherits everything the user has already told ChatGPT, so product recommendations can be grounded in the real conversation, not a cold search
• Deeper catalog access - the app can query the full product catalog via its own backend, not just the subset that made it into the ChatGPT Shopping feed
• Checkout continuity - combined with ACP and Instant Checkout, the user browses, selects, and pays without leaving ChatGPT

For retailers at the Enterprise tier, a ChatGPT App is the difference between being one of thousands of brands competing for a text mention and being a named, branded experience inside the conversation. It is the agentic-commerce equivalent of owning a storefront versus renting a shelf.

At the protocol level, a ChatGPT App is an MCP server with an extension for rich UI rendering. The developer defines:

Tools - the actions the app can take ("search products", "add to cart", "check inventory", "book appointment"). Each tool has a schema defining its inputs and outputs.

Resources - the data the app can provide (product catalogs, user history, booking availability). Resources are retrieved on demand.

UI components - the Apps SDK extends MCP with a rendering layer. The app can return structured UI payloads that ChatGPT renders inline - product cards, carousels, forms, confirmation dialogs, or full-screen views.

When a user converses with ChatGPT, the model decides when a registered app is relevant based on the user's intent and the app's declared capabilities. The model calls the app's tools, receives structured data, and renders the UI - all inline inside the conversation. The user can interact with the UI (click a product, submit a form) and the app handles the response.

For retailers, the typical App architecture connects to the same catalog, inventory, and checkout systems that power the main storefront. The App is a new surface for the same data, not a new data source. This is where a commerce infrastructure platform matters - a retailer with a single, normalized product catalog can power an App, a Google Merchant feed, and a traditional storefront from one source of truth instead of maintaining three separate integrations.