Too many requests

I think we need better error handling here and have to add warnings to the status page when an account does not fill the requirements of most 3rd party module. Only sparse usage of gpt-3.5 is allowed more or less.

As for support request - Could you verify that you have added $5 to the account - that is the limit to reach tier1. If you have this, could you verify under https://platform.openai.com/settings/organization/limits that you are tier 1.

Tier 1 should in general be enough, to handle most rate limits for any website where only administrators/editors has access to the features and use them. For end-user facing solutions you have to have a higher tier.

More about tiers here: https://platform.openai.com/docs/guides/rate-limits/usage-tiers?context=...

If you are on Tier 1, please let us know and we have to look into what is not correct.

Hi Marcus

The paid account did solve the Too Many Requests issue once it's above Tier 1.
Thanks for pointing this out. Maybe there should be a documentation available somewhere explaining this, or I might have looked over it? :-)

The chat now works, but now I'm still stuck when trying to translate.
I keep getting a "The Prompt is unsafe: The prompt was flagged by the moderation model." error when translating. Might be related, but maybe something for auto_translation? Any ideas?

Just to answer your question - we have OpenAI moderation endpoint running on every request by default. You can turn this off by configuration, but its really highly recommended that you do not.

If you get enough strikes of calls against the LLM that are flagged as suspicous, they might ban your account.

So the error you saw, was because the article flagged something in the content as harmful, see: https://platform.openai.com/docs/guides/moderation

If you want to try to figure out what it is there is a submodule called AI API Explorer you can enable, where you can go to the moderation explorer and paste in your article and you will get a verbose result with what type of harmful category flagged it.

I added this issue: ✨ Improve checks for tier 1 accounts Active

Could this issue be closed as answered?

Thanks for the explanation.
This can be closed for me.