Acceptable Use Policy

1. Scope and responsibility

This Acceptable Use Policy (“AUP”) applies to every user, customer, agent, application, integration, and workload that accesses Neutron AI. Capitalised terms have the meanings in the Terms of Service. Customer is responsible for its Authorised Users, agents, applications, Customer Content, tools, and downstream use of Output.

You must use the Service lawfully, safely, in good faith, within documented technical limits, and with safeguards proportionate to the sensitivity and impact of the workload. You may not help another person evade this AUP.

2. Illegal, deceptive, or harmful activity

You must not use the Service to create, facilitate, promote, conceal, or materially assist:

• activity that violates applicable law, regulation, court order, sanctions, or another person’s legal rights;
• fraud, scams, impersonation, deceptive practices, credential theft, phishing, money laundering, or evasion of legal or safety controls;
• violence, terrorism, violent extremism, credible threats, targeted harassment, stalking, or instructions whose primary purpose is to cause physical harm;
• development, acquisition, or operational use of biological, chemical, radiological, nuclear, or other weapons, or meaningful assistance that increases the ability to cause mass harm;
• unlawful gambling, trafficking, controlled-substance distribution, or other organised criminal activity; or
• false claims that an Output, predicted consequence, or AI-generated statement is verified, guaranteed, professionally approved, or an objectively optimal decision when it is not.

3. Security and service integrity

You must not:

• probe, scan, exploit, disrupt, overload, or gain unauthorised access to the Service, another tenant, account, Nucleus, Scope, system, or network;
• upload malware, destructive code, unauthorised cryptominers, denial-of-service traffic, or payloads intended to escape isolation or manipulate infrastructure;
• circumvent authentication, permissions, rate limits, usage metering, safety filters, deletion tombstones, retention controls, approvals, or technical restrictions;
• share, publish, sell, or expose credentials, or use credentials issued to another workspace without express authority;
• use automated means to scrape, enumerate, or bulk extract non-public Service data outside documented interfaces;
• benchmark or security-test the Service in a way that risks harm or public disclosure without written authorisation; or
• use the Service to attack, compromise, or evade protections of a third-party system.

Good-faith security research should be reported privately to admin@neutronai.dev. Do not access data beyond what is necessary to demonstrate an issue, degrade the Service, or publicly disclose an unresolved vulnerability.

4. Privacy, surveillance, and sensitive data

You must not use the Service to:

• collect, infer, store, retrieve, disclose, or profile personal data without a valid lawful basis, required notice, and appropriate rights process;
• conduct unlawful or indiscriminate surveillance, track a person without authority, identify an anonymous person, or infer highly sensitive traits in a manner prohibited by law;
• build or enrich facial-recognition, biometric-identification, social-scoring, or location-tracking databases in violation of law or without required safeguards;
• submit passwords, private keys, bearer tokens, session cookies, raw payment-card data, or other live authentication secrets as memory content;
• submit health, biometric, genetic, precise-location, criminal-offence, financial-account, children’s, or other highly sensitive data unless your contract and configuration support it and you have completed required assessments; or
• defeat a valid access, deletion, restriction, objection, retention, consent, or opt-out instruction.

You must minimise data, isolate it by validated Nucleus and Scope, restrict access, set an appropriate retention period, and delete it when no longer needed.

5. Safety, exploitation, and rights of people

You must not use the Service for:

• child sexual abuse material, grooming, sexual exploitation, age-inappropriate sexual content involving minors, or any system that facilitates harm to a child;
• non-consensual intimate imagery, sexual deepfakes, sextortion, or sexual content involving a person without required consent;
• targeted harassment, bullying, doxxing, coercive control, or exploitation of vulnerability;
• discrimination prohibited by law or decisions based on protected characteristics where the use is unlawful or lacks required safeguards;
• manipulation that materially impairs an individual’s ability to make an informed decision and is likely to cause significant harm; or
• deceptive impersonation of a real person, organisation, public authority, or trusted professional.

Apparent child sexual abuse material or credible threats may be preserved and reported to appropriate authorities where law requires or permits.

6. High-impact and regulated decisions

You may not use Neutron as the sole or determinative basis for a decision that produces legal or similarly significant effects in employment, worker management, housing, education admissions, credit, lending, insurance, healthcare, legal services, migration, policing, criminal justice, access to essential public services, or another high-impact domain unless all of the following are true:

• the use is lawful and your plan, Order, and Documentation expressly support it;
• a qualified human meaningfully reviews the relevant evidence, uncertainty, alternatives, and Output before the decision;
• you have completed required risk, data-protection, equality, fundamental-rights, and algorithmic impact assessments;
• affected people receive required notice, explanation, contestability, accessibility, and human-review rights;
• you test and monitor accuracy, bias, security, drift, and foreseeable misuse using representative evidence; and
• you retain an auditable record of inputs, approvals, policy, outcome, and recovery path without storing hidden chain-of-thought.

Neutron Output is decision support and is not medical, legal, financial, employment, safety, or other professional advice. A qualified professional must make and own any regulated decision.

7. AI agents, consequence simulations, and external execution

You must disclose AI interaction or synthetic content where required, preserve meaningful human control, and avoid representing probabilistic Output as certain. Depth, branch, model-call, runtime, node, cost, and scenario budgets must remain bounded. Consequence simulations at depth 50 or more must use supported asynchronous and cancellable paths; depth 1000 requires explicit confirmation.

High-risk or irreversible actions require a recorded human approval. External execution additionally requires an authorised, scoped, idempotent action adapter; explicit tool policy and permissions; validated inputs; recovery and rollback planning; audit evidence; and confirmation immediately before execution where appropriate. Do not use memory or generated Output to bypass an approval gate.

You must not request, persist, expose, or attempt to extract raw chain-of-thought, hidden prompts, model scratchpads, or private reasoning. Use concise explanations and structured, reviewable artifacts instead. Reflection lessons from sensitive domains require review before memory write-back.

8. Content and intellectual-property rights

You must not:

• submit or use content you do not have the right to process;
• infringe copyright, trademark, database, privacy, publicity, confidentiality, trade-secret, or other rights;
• remove provenance, ownership, safety, or synthetic-content notices in a deceptive or unlawful way;
• use the Service primarily to reproduce, reconstruct, or distribute protected material unlawfully; or
• misrepresent generated, retrieved, or transformed content as a human-authored or authoritative source where disclosure is required.

9. Communications, spam, and resource abuse

You must not use Neutron to send spam, unlawful marketing, phishing, deceptive bulk messages, or communications that violate consent and opt-out rules. Do not create excessive requests, accounts, Nuclei, Scopes, or workloads intended to avoid limits, degrade service, impose unreasonable cost, or deny capacity to others. You must respect documented rate, storage, retention, and concurrency limits and cooperate with reasonable abuse-prevention controls.

10. Enforcement, cooperation, and reporting

We may investigate suspected violations and may restrict content, revoke credentials, rate-limit, suspend, or terminate access as reasonably necessary to prevent harm, comply with law, or protect the Service and others. Where appropriate, we will consider severity, intent, history, actual or potential harm, and whether the issue can be safely cured. We will ordinarily give notice and an opportunity to respond unless doing so would increase risk, interfere with an investigation, or be legally prohibited.

We may preserve relevant records and cooperate with lawful authorities. Enforcement does not require us to monitor all Customer Content and does not shift Customer’s responsibility for its systems or users. To report abuse, a safety issue, or a suspected violation, email admin@neutronai.dev with enough non-sensitive detail for us to assess it. Do not include credentials or unnecessary personal data.

Contact and company information

Questions about this document may be sent to admin@neutronai.dev.