Security and Privacy
MORATUS doesn't run its own AI model; replies are generated with models from third-party model providers. When you send a message, the content needed to generate the reply is sent to these providers. Your inputs and outputs belong to you.
What is sent?
- Your message and the content of the files you attach.
- The chat history needed to keep the reply consistent.
- Relevant memory records, if memory is on.
- Content fetched from connectors or web search for the task, and the instructions of scheduled tasks.
- Your account and payment details aren't sent unless they're needed to generate the reply.
Where, and to how many providers?
- Depending on the model you choose and on routing, a request can go to at least one and at most three providers; for example, the steps of a task can run on different models, or another provider can step in when one can't be reached.
- Providers use infrastructure in the European Union (Germany), Singapore, and China. In exceptional cases where none of the primary providers can be reached, a request can be sent through a model routing service in the US.
Retention and model training
- MORATUS doesn't use your content to train models on its own behalf; if that changes, your explicit consent is asked for first.
- Your content isn't sold or used for advertising.
- Providers' terms differ: some don't store API content, some keep it for up to 30 days for abuse monitoring, and some providers' terms may allow content to be used to improve their own services.
- Incognito and guest chats aren't saved in MORATUS; to generate replies, those messages are still sent to the provider.
Recommendation
Unless the task requires it, don't enter special categories of personal data such as health, biometric, or belief data, other people's personal data, trade secrets, or information under a confidentiality obligation into MORATUS.
You can apply under KVKK to learn which providers your data has been sent to. Details: Privacy Policy: AI model providers and International data transfers.