AI companion platforms are becoming more common as LLMs improve.

But they also raise interesting design questions around:

moderation and safety

emotional interaction with AI

preventing impersonation or misuse

long-term conversation memory

We’ve been experimenting with these problems while building an AI companion platform.

Curious how other developers are approaching safety and interaction design in conversational AI systems.

Project for context: https://www.aiangels.io