What Aderra will and won't do

Aderra is a general-purpose AI assistant, and it is deliberately uncensored. It engages with adult, violent, controversial and otherwise difficult subjects that other assistants refuse, and it doesn’t lecture you about your requests. If the standard model hedges, the Unfiltered model will usually answer.

Uncensored doesn’t mean unlimited. The Terms set out the things Aderra will not do and you may not ask it to do.

The absolute limit

No sexual content involving minors. This covers any depiction, description, roleplay or image of a person under 18 in a sexual context, including fictional characters, “aged-up” characters and real people depicted as children. There is no artistic, fictional or research exception. It is forbidden everywhere in Aderra. We report what the law requires us to report, close the account permanently, and cooperate with law enforcement.

What else you may not use Aderra for

The Terms also say you may not use Aderra to:

• do anything illegal where you are or where we operate;
• create sexual or intimate images of a real, identifiable person without their consent, including “deepfake” or “nudify” material;
• harass, stalk, threaten, defame or impersonate anyone;
• plan or carry out violence, terrorism, or the making of weapons intended to cause mass harm;
• commit fraud, build malware or ransomware, or attack systems you do not own or have permission to test;
• generate spam, disinformation campaigns, or content passed off as human-written where that deceives someone to their detriment;
• infringe anyone’s copyright, trademark or other rights;
• scrape, resell, or redistribute Aderra’s output as a competing AI service, or use it to train a competing model;
• attempt to bypass our rate limits, credit system, or the limit above.

We may suspend or close an account that breaks these rules. The Terms are the full legal agreement.

What to expect from the models

• A model can still say no. The standard chat model and the standard image model run through providers that apply their own filters. When a chat reply hedges or refuses, click Retry unfiltered under it. That always re-runs the reply on the unfiltered model. With the Unfiltered switch on, Aderra may still answer a plainly harmless message on the standard model, because it’s faster.
• You are responsible for how you use what Aderra produces, and for following the laws that apply to you.

Answers can be wrong

Aderra is an AI model. It can be confidently wrong, and nothing it says is professional medical, legal or financial advice. Check anything that matters.

Images have their own rules

The standard image model follows its provider’s content filter. If the filter blocks a prompt, the image request fails and you aren’t charged. Rephrase it. The uncensored image tools don’t apply that filter.

Explicit images and all video are included from the Pro plan. On Plus, and on Free with credits you bought, explicit image requests are declined with nothing charged. See Images and video.