FAQ
Questions, answered plainly.
If yours isn't here, write to wave@waveai.cloud and a person will answer it.
Getting started
What is Wave AI?
An independent AI safety and development lab. We build our own models, red-team them hard, and release only what holds up. Our first model, Wave-1, is live now.
Is Wave-1 free?
Yes. Chatting is free, with fair-use limits that keep it fast for everyone. If you hit the limit, it resets on its own.
Do I need an account?
No. Open the chat and start typing. Nothing to sign up for, nothing to install, no card.
Does it work on my phone?
Yes. The chat is built for phones as much as desktops, in any modern browser.
Which languages does it speak?
The whole experience is available in English and Arabic, right-to-left layout included, and you can chat in either.
The model
What is Wave-1 good at?
Everyday thinking work: writing and rewriting, code, math, explanations, planning. For its size it is strongest on knowledge and math, and it tells you when it is not sure. See the benchmarks.
Can it search the web?
Yes. Turn on Search in the composer and Wave-1 looks things up before answering, with sources you can check under the reply.
What is Memory?
Notes Wave-1 keeps in mind in every chat, like "I prefer short answers." Add them from the sidebar or just say "remember that...". They stay in your browser and you can delete them anytime.
Does it make mistakes?
Sometimes, like every model. It is built to admit uncertainty instead of hiding it. Check important information, especially anything you would act on.
What does the Experimental badge mean?
Wave-1 is our first release and we improve it often, so its capabilities keep moving. The safety bar is the part that never does.
Privacy and safety
Where are my chats stored?
In your browser, on your device. Clearing your browser data clears them. Messages are sent to our model service to generate each reply; we don't sell data and we don't run ads. Details in the privacy policy.
What will Wave-1 refuse?
Anything built to cause harm, however it is phrased, encoded, or roleplayed. Every message passes Wave Guard before the model answers. Ordinary questions go straight through.
Why did my chat get locked?
A request in it crossed a safety line, so that conversation is closed. Start a new chat with one click; nothing else is affected. If you think it was a mistake, tell us at wave@waveai.cloud.
How do I report a problem?
Email wave@waveai.cloud with what you asked and what it answered. Safety reports go straight into our next red-team run, and both kinds matter: things that got through, and ordinary requests wrongly refused.
Still curious? Ask the model.
Wave-1 answers questions about itself too, in plain language.