Acceptable Use Policy

1. Purpose

This Acceptable Use Policy ("AUP") defines prohibited uses of the Sentiqra API and platform. It forms part of the Terms of Service. Sentiqra provides semantic-analysis signals; customers remain responsible for how those signals are used.

2. Prohibited uses

You may not use Sentiqra to:

• Violate any applicable law, regulation, or third-party right
• Process content you are not authorized to process
• Submit content that applicable law prohibits you from possessing or transmitting. This clause does not prohibit safety analysis itself: submitting harmful, abusive, or suspected unlawful material for detection, moderation, compliance, or legitimate research is the intended use of the service - you remain responsible for ensuring your specific submission and downstream use are lawful.
• Build profiles of, surveil, or make automated decisions about identifiable individuals where doing so is unlawful
• Generate, facilitate, or amplify illegal content, malware, fraud, or social-engineering campaigns
• Attempt to reverse-engineer, extract, or replicate the detection models or taxonomies
• Circumvent rate limits, authentication, or access controls
• Resell or redistribute the service without a written agreement
• Interfere with or disrupt the service or its infrastructure

3. Deployment assessment and intended use

Deployments of restricted or sensitive subjects require a recorded intended-use assessment before access is granted. For each proposed deployment we record the input modality, the people affected, the decision context, the consequences, human involvement, jurisdictions, and each party's responsibilities. Counsel then identifies applicable restrictions, assessments, notices, and safeguards, and distinguishes intended uses from prohibited or unsupported uses.

Assessment is based on the deployment's actual inputs and use - not on the product or analyzer name. A non-generative system, a "signals only" API, or a no-training policy does not by itself make a downstream application acceptable. Where outputs are linked to identifiable people, or sensitive categories or consequential decisions are involved, additional review (for example under GDPR or the EU AI Act) may apply regardless of what the analyzer is called.

Approvals are scoped. A materially changed use case - different inputs, different affected people, different decisions or consequences - requires a new assessment and cannot rely on a prior approval issued for a different scope.

4. Signals and consequential decisions

Analyzer signals are decision-support inputs - match-strength indicators, not verdicts. Supported workflow patterns include routing, triage, prioritization, analyst review queues, warnings and friction, appeals replay, and aggregate analytics.

Automatic adverse action against users, accounts, transactions, or content - blocking, denying, removing, restricting, or suspending - is permitted only where all of the following hold:

• Validated policy - the acting threshold or rule was measured on your own traffic for that specific use case, including benign look-alike content
• Human review - a designated reviewer handles consequential decisions; signals escalate borderline or contested cases rather than silently resolving them
• Appeal path - affected users can contest the action and reach a human decision

Until those conditions are met for a given use case, automatic adverse action must stay disabled and signals remain review-only. For analyzers, languages, or content types whose look-alike false-positive profile has not been measured on your traffic, signals are review-only inputs.

5. Reporting violations

To report a violation of this policy or abuse involving Sentiqra, use the Report Abuse page or email abuse@sentiqra.tech. For security vulnerabilities, see Report a Security Issue.

6. Enforcement

We investigate reported violations and may take action proportionate to severity:

• Warning - for first or minor violations
• Suspension - for serious or repeated violations, pending investigation
• Termination - for severe, unlawful, or persistent violations

We may suspend access immediately where required to prevent harm, illegal activity, or risk to the service.

7. Appeals

If your access is suspended or terminated and you believe the decision was in error, contact legal@sentiqra.tech within 30 days to request a review.

8. Changes

We may update this policy. Material changes are notified via email or a notice on the website. The "Last updated" date reflects the most recent revision.