OGAN AI Blog & Media

How to Create a Custom GPT Agent for Your OGAN AI Project?

2025-07-16
You can now create your own GPT, fully tailored to your project, directly within ChatGPT by connecting Your GPT to OGAN AI's MCP Server. This feature is currently available only to ChatGPT Plus members.

To get started, go to the GPT Editor and follow these steps:
First, use the instructions template to configure your GPT.
This setup allows the agent to send the command and text fields from the user's input directly to your MCP API-and return the response accordingly.

Next, you will need to handle authentication.
When creating a new action, make sure to enter your Bearer token.
To get this token, just log in to ogan.ai, select your project-or create one if you haven't already-then run the /mcptoken command in the chat agent. You’ll receive your token there. Use it as a Bearer token in the Authorization field.

To speed up development, you can import your API definition using the provided OpenAPI YAML file URL.
Want to make your GPT even smarter?
Just upload the command reference file - ogan_ai_command_list.txt - to give your agent full context on what it can do.
And for the best performance, always use your Ogan AI chat agent directly on ogan.ai.
It's constantly updated and already knows the structure of your database.

Important reminder:
When you publish your GPT, select "Only me" so it's private.
Do not publish it to the GPT Store or share it via public link—because your token gives access to your project, and you want to keep that secure.

And that is it!
You are now ready to manage your entire project and interact with your live database through your own personalized GPT agent.
You can also find other helpful resources-like the plugin JSON files, guides in MCP authorization blog content on the ogan.ai website.
https://ogan.ai/blogcontent.php/mcp-authorization-27