ORBINTU — AI + Data Terms
Version 1.0 · Draft — not yet in effect · Last updated 26 SEPTEMBER 2026
https://orbintu.com/legal/ai · hello@yuvi.it.com
In short
- ORBI understands and forms work. NTU executes it. Neither gains authority just by having context.
- AI output can be wrong. Verification, not the claim of completion, closes a Need.
- Consequential actions require the permission you granted and, where marked, your approval.
- ORBINTU does not train AI models on your content, and no AI provider receives your content at this release.
This summary is for convenience. The numbered sections below are the full text.
01Current status
ORBI and NTU are not yet available as a live service. The demonstrations on this website use a fixed, scripted scenario and do not process your personal information. These terms describe how the service is designed to work, and will apply when AI features launch.
02The system model
ORBI is the persistent intelligence system that develops a working model of a person and their world, interprets what they are trying to accomplish, determines what is missing between current and desired reality, assembles relevant information and intelligence, and forms correct work for NTU.
NTU is the Navigating Task Unit. It receives correctly formed, bounded work from ORBI and navigates the capabilities required to execute it without losing the work’s purpose, permissions, or success conditions.
In short: ORBI understands and forms the work, NTU carries out the machine side of it under the authority attached to that work, and ORBI checks the result.
03What context ORBI may process
ORBI may process the information you give it directly, your ORBI profile, and data from Connected Services you have authorised, within the scope of each authorisation. ORBI does not gain permission to act merely because it has context.
04What NTU receives
NTU receives bounded work: the purpose of the work, the permissions attached to it, and the conditions for success. It is designed to receive only the context needed for that task, and to operate within the task’s defined bounds.
05Third-party AI and model providers
No third-party AI or model provider receives user prompts, files or other User Content through the ORBINTU service at this release.
When an AI feature that sends User Content to a provider is enabled, we will name the provider and describe its data-use terms in the Privacy Policy and in these terms before the feature is activated.
06Connected applications
Each Connected Service has its own access scope, granted by you and revocable by you. NTU acts inside a Connected Service only through the scope you granted, and only for work that carries the necessary authority.
07Use of your content for model training
ORBINTU does not use User Content to train AI models.
We use content to improve the product only when you explicitly choose to send it to us as feedback.
08Retention
Prompts, work, outputs and evidence are not collected at this release, because no AI execution service is active. A retention policy for them will be set and published here before an AI feature collects them.
09Human review
ORBINTU staff may review material you provide only:
- when you explicitly ask us for support with it;
- when needed to investigate security, abuse or legal issues;
- where otherwise required by law.
Access is limited to the people who need it, and is logged where our systems support it.
10Sensitive information
Avoid giving ORBI sensitive information it does not need, such as government identifiers, payment card numbers, passwords, or health information, unless a feature is specifically designed for it.
12Output accuracy and verification
AI output can be incomplete, inaccurate, or out of date. It is not professional advice.
NTU reporting that a task is complete is evidence, not proof. ORBI verifies results against the work’s success conditions. Verification, not the report of completion, closes a Need. Verification can also be wrong, so review consequential results yourself.
13Your responsibilities
Review output before relying on it, grant only the permissions you intend, and follow the Acceptable Use Policy. You remain responsible for actions you approve.
QUESTIONS
hello@yuvi.it.com