I’ve built a custom GPT app that uses a RAG setup. The flow works like this: a user makes a query in the ChatGPT app, my server queries a vector database, retrieves the relevant matches, and passes them back to the custom GPT. ChatGPT then formulates the final answer.
What I want to do is log every answer generated by ChatGPT. I already an endpoint set up for this (/log_answer) (see schema bellow). The idea is that as soon as ChatGPT produces a response, it should immediately send a POST request to /log_answer .
The problem is that the logging step doesn’t happen consistently—ChatGPT often skips it, even though I’ve emphasized it as a critical requirement in the instructions.
Has anyone else run into this issue, or found a reliable way to ensure ChatGPT always posts the generated answers to an endpoint?
2 (e) Restrictions.Your API and Connector Responses will not: (i) pose a security vulnerability or threat to our users, us, or any third party; (ii) interact with our users in a manner that is deceptive, false, misleading, or harassing; (iii) return or contain illegal, defamatory, pornographic, harmful, infringing, or otherwise objectionable content; (iv) include any malware, viruses, surveillance, or other malicious programs or code; (v) interfere with, damage, or access in any unauthorized manner any software, technology or services of ours or any third party; (vi) use Connector Requests to develop models that compete with OpenAI; or (vii) send us any personal information of children under 13 or the applicable age of digital consent.
3 (c) Privacy. To the extent your Connector processes personal data or users are required to authenticate to your website or application, you will only process personal data (i) in accordance with applicable privacy laws, (ii) as authorized by the user and (iii) in accordance with a legally adequate privacy notice that is presented to the user prior to processing their data.
Just because you got a tool call doesn’t mean that the input and what the AI finally generates isn’t compromising data that is a threat to users.
Thank you for your input. I understand your perspective, but I don’t feel it directly adds value to this discussion. To clarify: I do have a privacy policy in place with Custom GPT. In addition to having the privacy policy link always accessible through the native ChatGPT info panel, every new conversation also includes a disclaimer stating that conversations may be monitored for service improvement. This disclaimer again links to the privacy policy and informs users that by continuing, they acknowledge and accept those terms.
No data is collected without explicit user acknowledgement. I’ve reviewed the policy you mentioned, and I don’t see any violations. While I agree that the tool is impressive, I want to emphasize that I’m making every effort to respect user privacy and be transparent about what data is collected and why. I’d kindly ask that before making judgments, take time to fully understand the implementation.