ChatGPT does not support the Model Context Protocol natively as of mid-2026. Claude Desktop has built-in MCP support; ChatGPT does not. However, you can achieve the same outcome — querying live Google Ads data and executing approved changes through conversation — using ChatGPT's Custom GPT Actions. The setup is more involved than Claude's native MCP, but the end result is comparable: a ChatGPT interface that calls Digital Face's API and works with your actual account data.
This guide assumes you have a ChatGPT Plus or Team subscription (Custom GPT creation requires a paid plan) and a Digital Face account with at least one ad platform connected. The full setup takes 20 to 30 minutes.
Step one: set up Digital Face. Sign up at digital-face.nl/signup and connect your Google Ads or Meta Ads account via OAuth. Once connected, go to Settings and open the API Access section. Create an API key with read and write permissions. Copy the key — you will need it in step four.
Step two: go to chat.openai.com and open the GPT builder by clicking Explore GPTs, then Create. Give your GPT a name like 'Ad Account Manager' and a description. In the Instructions field, describe what this GPT does: it connects to your ad management platform to provide live campaign data, performance analysis, and the ability to execute approved changes.
Step three: click Configure, then scroll to Actions and click Create new action. This is where you define what API endpoints ChatGPT can call. In your Digital Face dashboard, go to Settings → API → OpenAPI Schema and download the JSON schema file. Paste the content into the Schema field in the Action editor. This schema tells ChatGPT which Digital Face tools are available and what parameters each expects.
Step four: configure authentication. In the Action editor, set Authentication Type to API Key. Enter your Digital Face API key from step one. Set the header name to Authorization and the format to Bearer Token. This authenticates every request ChatGPT sends to the Digital Face API.
Step five: test the integration. Save your Custom GPT and start a conversation. Try: 'What campaigns are currently active in my Google Ads account?' ChatGPT will call the Digital Face API, receive structured data, and return a plain-language summary. If you see an error, the most common cause is an incorrect API key or a schema field that does not match the current API version — download the latest schema from your Digital Face dashboard and re-paste it.
The practical difference between ChatGPT Custom Actions and Claude's native MCP is setup friction, not capability. Claude's MCP support is built in — add a config file and restart. ChatGPT requires building a Custom GPT and configuring an OpenAPI schema. Once both are set up, the conversational experience is similar: ask a question, receive an answer grounded in live account data, execute approved changes from the chat interface.
If your team already uses ChatGPT for copywriting, strategy, and client communication, adding ad account access via Custom Actions keeps everything in one interface. If you are starting fresh and choosing an AI client specifically for ad management, Claude Desktop's native MCP support is the lower-friction path. Digital Face works with both. Connect your account at digital-face.nl and choose the client that fits your workflow.