Content Moderation Policy

Last updated: August 22, 2021

This Content Moderation Policy explains how we review and respond to content and conduct on services operated by Anthony Kung ("we", "us", or "our"). It is intended to provide a clearer understanding of the standards we apply, the kinds of content that may be restricted, and the enforcement tools we may use to keep the Services usable, safe, and legally compliant.

Content moderation necessarily involves judgment. Different contexts matter, and no policy can enumerate every possible scenario. For that reason, this policy describes broad categories of prohibited content and the principles that inform moderation decisions, rather than attempting to reduce all outcomes to a rigid checklist.

1. Scope

This policy applies to content made available through the Services, whether submitted directly by users, sent through integrations, generated through AI features, uploaded through APIs, or otherwise transmitted using our infrastructure.

Covered content may include:

text and documents
prompts and model inputs
generated outputs
images, files, and attachments
comments, feedback, and support submissions
metadata or structured content sent through APIs

2. Moderation Objectives

Our moderation practices are designed to balance user expression, platform usability, legal compliance, product integrity, and user safety. In practice, this means that we may take action not only against clearly unlawful content, but also against content or behavior that creates a meaningful risk of abuse, fraud, exploitation, or service disruption.

Moderation is also necessary because platform misuse can affect more than the person who posts the content. Abusive material can expose other users, degrade trust, consume infrastructure resources, and create operational or legal risk for the platform as a whole.

3. Prohibited Content and Conduct

Content or conduct may be removed, restricted, or escalated if it includes, facilitates, or promotes:

illegal content or unlawful activity
harassment, threats, hate speech, or targeted abuse
sexually exploitative content or content involving minors
malware, exploit payloads, phishing content, or credential theft
fraud, impersonation, scams, or deceptive conduct
non-consensual personal information, doxxing, or privacy violations
intellectual property infringement or misuse of confidential information
attempts to evade platform restrictions, safeguards, or moderation systems
spam, manipulation, or low-quality abusive automation

The same content may be assessed differently depending on context. For example, descriptive discussion of harmful topics for reporting, research, journalism, or safety analysis may not be treated the same as content intended to facilitate or celebrate harm.

4. Detection and Review Methods

We may use a combination of automated systems and human review to identify policy violations. Automated review may include heuristics, reputation systems, trust and safety models, spam detection, anomaly detection, keyword matching, or other operational safeguards.

Automated systems improve response speed and scale, but they are not perfect. They may generate false positives or false negatives. As a result, moderation decisions may be based on a mix of automated flags, user reports, manual investigation, and contextual review.

5. Enforcement Actions

Where we believe content or behavior violates this policy or otherwise threatens the Services, we may take one or more actions depending on severity, context, recurrence, and risk.

Possible actions include:

refusing to publish or process content
removing, disabling, or suppressing content
warning the user or requesting clarification
limiting features, visibility, or reach
rate-limiting API keys, workspaces, or accounts
suspending or terminating access to part or all of the Services
preserving evidence and cooperating with legal process where appropriate

We may act with or without prior notice where necessary to protect users, third parties, or platform integrity.

6. User Reports

Users may report suspected abuse, policy violations, or harmful content. Useful reports generally include enough context to identify the issue and evaluate the claim, such as links, timestamps, screenshots, request identifiers, account information, and a short explanation of the concern.

We may prioritize reports based on severity, credibility, available evidence, and potential harm. Not all reports will result in visible enforcement action, and we may be limited in the detail we can share about our review.

7. Appeals and Reconsideration

If you believe content was removed or restricted in error, you may contact us and request review. Appeals should identify the relevant content or account action and explain why you believe the decision was incorrect.

We may decline appeals that are abusive, repetitive, unsupported, or clearly inconsistent with this policy. Submission of an appeal does not guarantee reversal of the original decision.

8. Moderation Records and Data Handling

We may retain reports, moderator notes, enforcement history, request metadata, and related operational records for as long as reasonably necessary to investigate abuse, prevent repeat violations, maintain security, resolve disputes, or comply with law.

Additional information about our handling of personal data is provided in the Privacy Policy.

9. Changes to This Policy

We may update this policy as product capabilities, legal obligations, and trust and safety risks evolve.

10. Contact

Questions or moderation-related reports may be sent to support@anth.dev.