Content Moderation Policy

Last updated: 25 July 2026

1. What this page covers

Our Terms of Service (Section 5) sets out what content is prohibited on Defici. This page describes, in plain terms, how we currently detect and act on listings — automated screening plus human review. It does not replace or narrow the Terms: a category not covered by automated screening is still prohibited under the Terms, and may still be acted on by our team.

2. Automated screening at submission

Every listing is checked automatically at the moment it's submitted, before it goes live. Two independent checks run:

  • A rule-based check for specific prohibited categories: weapons and firearms, controlled substances, fake or forged documents, and counterfeit goods. A match here blocks the listing outright — it is rejected at submission and never goes live.
  • A combination of rule-based checks and an AI content classifier that looks for other patterns our Terms prohibit, such as investment solicitation, attempts to move payment off-platform, or redirecting to another marketplace. A listing is flagged if either check flags it.

Automated screening covers a defined, limited set of patterns — it does not attempt to detect everything the Terms prohibit. Categories such as adult content, stolen goods, hate speech, intellectual property violations, or malware and phishing links are not currently covered by this automated layer; they rely on the human review described below.

3. What happens when a listing is flagged

A listing flagged by the rule-based category check above gets a 72-hour window to be edited into compliance. We email the seller with the reason and the deadline; the listing stays unpublished during this window. If a compliant edit isn't submitted in time, the listing is automatically removed. If you believe a flag was made in error, contact [email protected] and a person will review it — there is currently no separate appeal process or ticketing system beyond this.

A listing flagged only by the AI content classifier (not the rule-based category check) stays unpublished but does not get an automatic 72-hour countdown — it is held for our team to review directly, as described in the next section.

4. Human review

Our team can remove any listing that violates the Terms, whether or not it was caught by automated screening — including listings in categories the automated checks above don't currently cover. There is no public "report this listing" button today; if you've seen something that violates our Terms, email [email protected] and it will be looked at.

5. What this policy is not

This page describes how moderation works today, not a guarantee of outcome or timing. It is not legal advice and does not modify the Terms of Service, which remain the governing document for what is and isn't permitted on Defici.