Setting up ChatGPT

Step 1: Access the Settings/Connectors interface

To access the interface to begin the integration, perform the following steps:

1. Log in to ChatGPT or the OpenAI Dashboard

2. Open your settings

   

3. Click Create app to navigate to your list of connectors and MCP servers
   

Step 2: Add a new MCP server

In ChatGPT or OpenAI Dashboard, add the new server:

1. In the Connectors / MCP section, click Add MCP Server (or similar action, depending on the platform you’re using).

   

2. Enter the appropriate information about the MCP server to define it.

   Field	Value
   Name	Human-friendly server label
   Description (optional)	Purpose of the MCP server
   Server URL	MCP endpoint (e.g. https://app.rfpio.com/oa/v1/mcp)
   Auth Type	Oauth

3. When you are finished, click Create. The server will now appear in your Connectors list.

Note: Please make sure the server is enabled. It cannot be used if it is not enabled.

Step 3: Activate your server

In ChatGPT or OpenAI Dashboard, activate the new server.

In ChatGPT or OpenAI Dashboard, users can now enter a prompt and get a response based on content in Responsive.

Using ChatGPT with Responsive AI

After your administrator has set up the integration, you can use ChatGPT with Responsive to access information from your Content Library, Information about projects you're working on, and more.

To connect ChatGPT with Responsive, perform the following steps:

1. In ChatGPT, click Apps.
   

   

2. Begin typing Responsive in the search bar.
   

   

3. Select Responsive from the app list and click Start chat.
   

   
   

4. When the chat window appears, you will see Responsive listed.
   

   

You can now enter queries about your projects, tasks, reviews, and more in Responsive.

If you have write and edit permissions for a project, you can tell Responsive AI to save generated answers to a project using the tool:

save_project_question_response

If a response already exists for the question, the tool will update it using the new information. The responses are saved as the configured field type:

• Full rich text: The answer will be full rich text with the same formatting. Character and word length limits will be enforced. If the text is too long, it’s rejected and you will get an error message

• Basic rich text: The answer will be basic rich text with the same formatting. If the generated text has formatting that is not supported, it will be changed. Character and word length limits will be enforced. If the text is too long, it will be rejected and you will get an error message 

• Dropdown: Only one answer will be saved. If you try to save more than one, you will get an error message and the response will not be saved

• Radio button: Only one answer will be saved. If you try to save more than one, you will get an error message and the response will not be saved

• Checkbox: Multiple answers will be saved, unless there are duplicates

You can tell ChatGPT where to find information to generate an answer from tags and collections.
For example, you can enter Where are your services hosted from tag Enterprise and the chatbot answers the question using only content with the Enterprise tag.