Y-Smart Chat FAQ
A short, honest guide to how Y-Smart Chat works so you can make informed choices every time you use it.
Quick Navigation
Treating Answers
How should we treat answers from Y-Smart Chat?
Like smart suggestions, not final truth. Models can be sharp and still be wrong or out of date. If it really matters, double-check with a trusted source. Enabling search on some models helps with recency, but it's not a guarantee.
Off-Track Responses
Why does it sometimes go off track?
AI can drift; this is called a "hallucination." Often, a simple retry or a clearer follow-up prompt brings it back. Treat the output as a starting point and verify anything important.
Model Training
Are our chats used to train your models?
No, we don't train our own models on your conversations. Where we can, we opt out of provider training as well. For supported providers, we configure models to avoid training on user inputs:
- OpenRouter: We block routing to providers that allow training on user inputs when that can be controlled.
Sometimes new models only work with provider-side training enabled. If we offer one, you'll see a clear warning and can opt out or pick another model per chat.
Using your own API key? Make sure your org settings disable training if that's your preference.
For more details, see our Privacy Policy.
Feature Requests
We want a feature. Will you build it?
We're actively improving Y-Smart Chat. If there's something you want, tell us. Good ideas move up the list.
Send requests to support@y-smart.chat.
Support
How do we get support?
If you have questions or need help, we're here. Our team handles support and aims to respond quickly.
Email: support@y-smart.chat
We typically respond within 24-48 hours. For urgent issues, include "URGENT" in your subject line.