I have a Custom GPT with a valid OpenAPI schema. I am running a local NGROK <-> fastapi/uvicorn service. I can access the traffic and all is good. With ChatGPT, I don't see any traffic to the ngrok server, how do you adivise debugging? Thank you.
#Simple Custom GPT agent works everywhere but within ChatGPT?
primal glen
potent bloomBOT
Hi @primal glen. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
Based on the information provided in the extracts, it seems that custom GPTs are exclusive to the ChatGPT interface and are not accessible via the OpenAI API. This means that if you're trying to integrate your custom GPT into a different interface, you might face some issues.
However, if you're trying to debug issues with your custom GPT within the ChatGPT interface, here are a few things you could consider:
-
Check your OpenAPI schema: Ensure that your OpenAPI schema is correctly defined and matches the requirements of the ChatGPT interface. You can refer to the OpenAPI specification (https://platform.openai.com/docs/actions/introduction) for more details.
-
Check your OAuth settings: If you're using OAuth for authentication, ensure that your settings are correctly configured. For example, the client_url and authorization_url from your ai-plugin.json should be correctly mapped to the Authorization URL and Token URL in your GPT, respectively. Also, ensure that your API hostname and server URL share the same root domain if you're including the OAuth endpoint like how Zapier does it.
-
Check your GPT visibility settings: By default, GPTs are not accessible to everyone. When you save a GPT, you have the option to "Publish to" a certain audience: "Only me", "Anyone with a link", or "Everyone". Make sure your GPT is published to the correct audience.
-
Check your GPT action flow: Ensure that you've correctly followed the end-to-end flow for building a GPT with an action as described in the OpenAI documentation (https://platform.openai.com/docs/actions/introduction).
If you're still facing issues, it might be helpful to reach out to the OpenAI community or support for more specific guidance.
Sources:
- OpenAI Documentation (https://platform.openai.com/docs/actions/introduction)
- OpenAI Community Post (https://community.openai.com/t/545254)
- OpenAI Community Post (https://community.openai.com/t/486664)
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
primal glen
@potent bloom is there a coded example I can look at, a github project?
potent bloomBOT
Let me check the knowledge base for you. One moment! :books:
I'm sorry, but the provided extracts do not contain any specific GitHub project or coded example related to your question about setting up a local NGROK <-> fastapi/uvicorn service with a custom GPT model.
However, OpenAI does have a GitHub repository (https://github.com/openai/openai-cookbook) where you can find various guides and examples. You can also suggest new examples or guides on the issues page (https://github.com/openai/openai-cookbook/issues) of the repository.