Disclaimer

Last updated 14 August 2026

The models can be wrong

Everything Chatxy shows you comes from an AI model, and AI models state false things with complete confidence. They invent citations, misremember dates, get arithmetic wrong and describe functions in software that do not exist. This is not an occasional glitch — it is how the technology behaves. Check anything that matters before you act on it.

Not professional advice

Nothing here is medical, legal, financial, tax or psychological advice, whatever the answer looks like and however confidently it is phrased. A model has no licence, no duty of care to you and no knowledge of your circumstances. For anything with real consequences, speak to a qualified professional.

We didn't build the models

Chatxy is an interface. The models belong to Anthropic, OpenAI, Google, DeepSeek, xAI, Moonshot, Alibaba and Perplexity. We are not affiliated with any of them and are not endorsed by them. Their names appear here to say which model you are talking to, and for no other reason. We cannot change what a model says, and we cannot promise it will behave the same way tomorrow — providers update and retire models on their own schedule.

Knowledge has a cutoff

Each model was trained on data up to some date and knows nothing after it. When you turn on web search it can look things up; when you don't, recent events are outside what it can know — and it may answer anyway rather than say so.

Generated images

Images are produced by generative models and may resemble existing work, contain artefacts, or be unsuitable for commercial use. Whether you can use one for a given purpose depends on the model that made it and the law where you are. That is yours to check.

Uploaded files

Text extracted from PDFs, Word, Excel and PowerPoint files is best-effort. Layout, tables, footnotes and images may be lost or reordered, so an answer about a document can be wrong because the extraction was, not because the model was. Long files are truncated to fit the context window, and the app says so when that happens.

Your use, your responsibility

You decide what to do with what you read here. We are not liable for decisions taken on the basis of a model's output. If that is not a risk you can carry for a particular question, this is not the right tool for that question.