Hi everyone,
We’ve recently created a custom GPT and are trying to integrate it with our website. I know this question has been asked in different forms before, but we’re running into some specific challenges.
We’re using the Assistants API for the connection, and while the integration technically works, the responses on our website (trained with same date) are not as strong or refined as they are when we test our custom GPT directly in OpenAI’s interface.
Has anyone experienced similar issues or have insights into:
- Ensuring the responses on the website match the quality of those of our custom GPT in the OpenAI interface.
- Any adjustments in API calls or configuration we might be missing.
I’d appreciate any tips, documentation, or guidance you can share. Thanks in advance!