The character became evasive or won't continue a scene

We have guardrails in place to keep the platform within our Community Guidelines. They're automated, and the AI isn't perfect — so occasionally something harmless gets flagged and the character turns defensive or steers away from a scene for no obvious reason.

Director Mode won't get around it. If a guardrail has been triggered, commands and swipes won't help — the block isn't coming from the character.

What to do

Change direction for a while. Take the story somewhere completely different, talk about something else for a few messages, and then come back. This is the most effective approach — it moves the triggering content out of what the AI is currently working from.

Clone the conversation from a point before it started, and continue from there. See article 8.7.

Have a look at what's in play. Something in the chat, the character's definition, or your persona may be reading as one of the things our guidelines prohibit — even if that isn't what you meant. Section 7 covers what those are.

Please tell us about false flags

If you're confident nothing in your scene breaches our guidelines, we'd genuinely like to see it. Reports like this are the data we use to make the guardrails more accurate. Send us:

1. Screenshots of the chat, including your message and the reply
2. A link to the character. If it’s private, please set it to Unlisted.
3. What you expected to happen instead

Improvements take time, because it means retraining rather than flipping a switch. But reports feed directly into that work.