Why the API is so censored?

I defined “block_none” for every category and the model still can’t say nothing. EVERYTHING is being censored. Nonsense.

Hi @bkz , Welcome to the forum.

Even when you set custom filters to block_none there are still internal safety filters running separately from your settings that could be blocking the query.

Thanks

3 Likes

Yes of course. But it shouldn’t as my prompts are 100% safe. This is an error.

Try your prompts in AI Studio - are they blocked there too? If so, you can report them as an error (assuming they are actually errors and not your definition of ‘censorship’ clashing with Google’s!)

1 Like

Do you mind sharing the prompt so I can test it?

Is this happening with all Gemini models?

1 Like