Moderation Policy
Every blink posted to Blinking Chorus goes through moderation before it appears on the wall. This page explains what we check, what we remove, who reviews, and how fast.
How moderation works
When you submit a blink, it's checked automatically in three layers:
- PII detection — Scanned for personal information like email addresses, phone numbers, and street addresses. Automatically rejected.
- Spam detection — Checked for repetitive patterns, promotional content, and link-heavy text. Automatically rejected.
- AI safety check — Sent to OpenAI's moderation API, which checks for hate, harassment, sexual content, violence, and self-harm. The text only — no personal data is sent.
What gets removed
Blinking Chorus is for small, honest moments. The following are never allowed:
- Hate speech — Content that attacks or demeans people based on race, ethnicity, religion, gender, sexual orientation, disability, or other protected characteristics
- Harassment — Targeted abuse, threats, or bullying directed at individuals or groups
- Sexual content — Explicit sexual material, regardless of whether it's described as consensual
- Violence — Glorification of violence, threats of physical harm, or graphic descriptions of violence
- Self-harm instructions — Content that encourages, instructs, or glorifies self-harm or suicide
- Doxxing — Sharing personal information about others (names, addresses, phone numbers)
- Spam — Promotional content, repetitive posts, or link-heavy text
- Child safety — Any content that endangers minors, including grooming or exploitation
Self-harm and difficult feelings
We take self-harm content seriously. Blinks that express difficult feelings but don't cross into instruction or glorification may still publish — with a resource overlay linking to support services. Blinks with high-risk self-harm content are held for human review before any decision is made.
If you're struggling, please reach out to someone who can help. Crisis lines are available in most countries. You are not alone.
Who reviews and how fast
The automated layers check every blink instantly — you'll know within seconds whether your blink is on the wall. A small percentage of borderline or high-risk content is flagged for human review by a moderator. Human review typically happens within 24 hours.
How to report a blink
If you see a blink on the wall that you believe shouldn't be there, you can report it using the menu (⋯) on any note card. Each report includes a reason and optional details. Reports are reviewed by a human moderator.
Reporting is anonymous. Your session information is not attached to the report.
Appeals and mistakes
Automated moderation isn't perfect. Sometimes a blink is held or rejected by mistake. If you believe your blink was handled incorrectly, you can email [email protected]. We review appeals case by case. There is no formal appeals process yet — this is a small project, and we do our best.
Changes to this policy
We may update this moderation policy as the project grows and our tools improve. We aim to be transparent about changes. This page will always reflect the current approach.