IsoKron

IsoKron Acceptable Use Policy

Effective date: [PENDING — set on launch] Last updated: 2026-05-09

This Acceptable Use Policy ("AUP") governs your use of the IsoKron platform at isokron.ai (the "Service"). It is incorporated into the Terms of Service. By using the Service, you agree to comply with this AUP and to ensure your authorized users do the same.

1. Prohibited Content and Use

You may not use the Service to develop, deliver, store, or transmit any of the following:

1.1 Illegal content and activity

  • Content that violates applicable law, including controlled substances, weapons trafficking, fraud, money laundering, sanctions evasion, or unlawful gambling.
  • Content infringing intellectual property rights, including unauthorized reproduction of copyrighted code, trade secrets, or trademarked material.
  • Content that violates export-control or sanctions law (including OFAC, EU, UK sanctions regimes).

1.2 Harm to people

  • Child sexual abuse material (CSAM) or any content that sexualizes, exploits, or endangers minors. We have zero tolerance and report to NCMEC and law enforcement as required.
  • Content that incites or facilitates violence against specific persons, groups, or institutions.
  • Content that promotes self-harm, suicide, or eating disorders.
  • Targeted harassment, doxxing, or stalking.

1.3 Weapons of mass destruction and high-risk dual-use

  • Instructions, schematics, or operational guidance for biological, chemical, nuclear, or radiological weapons.
  • Software designed to attack critical infrastructure (power grids, water systems, financial-clearing systems, hospital networks).
  • Cyber-weapons, ransomware, malware-as-a-service, or other software whose primary purpose is unauthorized intrusion or sabotage.

1.4 Security and platform abuse

  • Compilation of software designed to bypass authentication, encryption, or access controls of any system you do not own or have explicit authorization to test.
  • Compilation of phishing, credential-harvesting, or social-engineering toolkits.
  • Use of the Service to attack, probe, or attempt to circumvent IsoKron's security controls (including the Critic, tenant isolation, rate limits, BYOK protections).
  • Attempts to access another customer's Workspace, Output, or BYOK keys.
  • Use of automated tools to evade rate limits, generate synthetic load, or overwhelm the platform.

1.5 Deception, manipulation, and fraud

  • Compilation of systems designed to impersonate real individuals or organizations.
  • Compilation of systems designed to manipulate elections, public discourse, or financial markets through automated coordinated inauthentic behavior.
  • Compilation of systems designed to generate fraudulent reviews, testimonials, or social proof.
  • Compilation of high-risk financial advice, medical advice, or legal advice systems without appropriate human oversight, professional licensure, and regulatory compliance.

1.6 Privacy violations

  • Submission of personal data of any individual without legal basis (e.g., consent, contract, legitimate interest).
  • Use of the Service to perform unsolicited bulk-data scraping or aggregation in violation of Data Protection Laws.
  • Compilation of systems designed for covert surveillance or to enable stalking.

1.7 Adult content and explicit material

  • Sexually explicit content, regardless of consent or legality, except as may be expressly permitted in a future tier with documented age-verification and content-moderation requirements; not permitted on the v1 platform.
  • Content depicting non-consensual sexual situations.

1.8 Hateful content

  • Content that promotes hatred, violence, or discrimination on the basis of race, ethnicity, national origin, religion, disability, age, sexual orientation, gender identity, or veteran status.

2. Use limits and platform integrity

You may not:

  • Exceed published rate limits or attempt to circumvent them via automated retry, multiple workspaces, or other means.
  • Attempt to discover or exploit vulnerabilities in the Service except under a formal coordinated-disclosure program. Report security issues to security@isokron.ai.
  • Use the Service to develop, train, or improve a product that competes with IsoKron's compilation pipeline or Critic.
  • Reverse-engineer the Critic or attempt to elicit its rejection rationale through repeated probing. The Critic is designed to give a single generic rejection signal precisely to prevent iterative oracle attacks; treating that signal as a starting point for circumvention is itself a violation.
  • Modify, distribute, sub-license, or commercially exploit the Service or its Output in a way inconsistent with the Terms of Service.

3. Customer responsibilities

You are responsible for:

  • Reviewing every compilation Output before relying on it. The compiler produces structured artifacts intended as a starting point for engineering work, not a final deliverable. Output may contain errors, omissions, or content unsuitable for production.
  • Validating that Output complies with applicable law, your industry-specific regulations, and any third-party rights.
  • Securing your account credentials and BYOK keys. If a key is compromised, revoke it immediately via the dashboard.
  • Promptly reporting suspected violations of this AUP that you observe by other parties or in your own Workspace.
  • Communicating this AUP to all users you invite to your Workspace.

4. Enforcement

We may take any of the following actions in response to a violation, in our reasonable discretion:

  1. Issue a warning.
  2. Suspend specific Workspace activities (compilation, BYOK invocation).
  3. Suspend Workspace access pending investigation.
  4. Terminate the Workspace and delete its content.
  5. Report the violation to law enforcement, NCMEC, or other authorities as required by law.
  6. Cooperate fully with legal process.

We will generally provide notice and an opportunity to remediate when feasible and lawful, but we may act immediately and without notice when:

  • A violation poses an immediate risk to safety, security, or platform integrity.
  • A violation involves CSAM, terrorism, or other content for which we have legal reporting obligations.
  • The Critic detects a pattern of compilations indicating malicious intent.
  • We are required to act by law, court order, or regulatory authority.

5. The Critic and pre-compile filtering

Every compilation passes through the Layer 4 LLM-as-Critic egress filter before commit. The Critic's purpose is to detect:

  • Prompt-injection attempts.
  • Parasitic-chaining patterns (where a compilation appears benign in isolation but composes with other compilations or external tool surfaces to enable attack).
  • Compilations whose Output would itself violate this AUP.

The Critic returns one of three verdicts: clean (proceed), suspicious (proceed with operator review), or rejected (block commit). For "suspicious" or "rejected" verdicts, you receive a generic rejection message; we do not disclose the Critic's specific rationale, because doing so would convert the Critic into an iterative oracle for adversaries to refine attack inputs.

If you believe a compilation was wrongly rejected, contact support@isokron.ai with the compilation ID. We will review and may permit re-submission with adjusted content.

6. Reporting violations

To report a violation of this AUP by another party or by content you have observed in your Workspace:

  • General AUP violations: abuse@isokron.ai
  • CSAM or child safety: report-csam@isokron.ai (we forward to NCMEC)
  • Security vulnerabilities: security@isokron.ai
  • Intellectual property infringement: dmca@isokron.ai

7. Changes

We may update this AUP. Material changes will be announced at least 30 days in advance via email or in-product notice; continued use after the effective date constitutes acceptance.

8. Contact

  • AUP questions: legal@isokron.ai
  • Reports: abuse@isokron.ai