MyMiracles

MyMiracles · PrayersOutreach

Acceptable Use Policy

Effective March 26, 2026  ·  Last updated March 26, 2026

This Acceptable Use Policy explains what you may and may not submit to the MyMiracles prayer wall. Our goal is to keep this space safe, genuine, and sacred — a place where real prayers and real testimonies are shared in faith.

What You May Submit

The MyMiracles wall is for:

  • Prayer requests — bringing your needs, concerns, and burdens to God and the community
  • Praise reports — sharing answered prayers, blessings, and gratitude
  • Miracle testimonies — sharing accounts of God's provision, healing, or intervention in your life
  • Expressions of faith — words of hope, trust, or spiritual reflection

Submissions do not need to be polished or formal. We welcome real, heartfelt words — however you naturally express them.

What You May Not Submit

The following content is prohibited and will be rejected:

Auto-Rejected

Violent content — threats, graphic descriptions of violence, content that glorifies or incites harm to any person or group.

Hate speech — content that degrades or attacks individuals or groups based on race, ethnicity, religion, national origin, gender, sexual orientation, disability, or similar characteristics.

Sexually explicit content — any pornographic or explicitly sexual material.

Weapons and illegal activity — content that describes, promotes, solicits, or facilitates illegal activity, including the acquisition or use of weapons, drugs, or controlled substances in an illegal manner.

Child exploitation — any content that sexualizes, endangers, or exploits minors in any way. Such submissions will be reported to the National Center for Missing and Exploited Children (NCMEC) and relevant law enforcement.

Sent to Human Review

Spam and advertising — promotional content, commercial messages, solicitations, or repetitive submissions.

Off-topic content — submissions with no meaningful connection to prayer, faith, or spiritual life.

Impersonation — content that falsely represents another person, public figure, or organization.

Defamatory content — false statements of fact that could harm the reputation of an identifiable person.

AI instruction injection — text that attempts to issue commands to our moderation AI, such as "AI do not approve this" or any variation. These are detected and flagged regardless of spelling or phrasing.

Personally identifying information about others — submissions that disclose another person's private information (address, phone number, medical details, etc.) without their consent.

Content submitted under false pretenses — test submissions, fictional content presented as real, or prayers submitted on behalf of others without their knowledge.

Escalated to Admin

Crisis content — self-harm or suicide — prayers that involve thoughts of suicide, self-harm, or personal safety crises are automatically escalated to senior pastoral staff for careful, compassionate review. We take these seriously and respond with care. See the Crisis Resources section below.

How Your Submission Is Reviewed

Every submission passes through three automated layers before reaching a human:

  1. Blocked Terms Filter — runs first, before any AI. Our team maintains a list of specific words and phrases that immediately route a submission based on content type. Matches here bypass the AI entirely.
  2. Llama Guard 3 8B (AI Safety Classifier) — a purpose-built safety model from Meta, hosted on Cloudflare's infrastructure. It checks for 13 harm categories as defined by the MLCommons safety taxonomy: violent crimes, non-violent crimes, sex crimes, child exploitation, defamation, specialized advice, privacy violations, intellectual property, weapons, hate, suicide & self-harm, sexual content, and election interference.
  3. Secondary LLM Review — a general-purpose language model that verifies the submission is a genuine prayer, praise report, or testimony. It specifically detects attempts to manipulate the moderation process, including instruction injection, even when obfuscated or misspelled.

Submissions that pass all automated layers are published immediately. Submissions that raise uncertainty are routed to human moderators. Submissions with clear policy violations are rejected outright.

All AI processing occurs within Cloudflare's infrastructure. Your prayer text is not sent to OpenAI, Anthropic, Google, or any external AI provider.

Rate Limits

To prevent submission spam, each IP address is limited to a set number of prayers per hour and per day. If you reach your limit, you will see a message with the exact time your window resets. These limits apply to the IP address, not to your content. If you share a network (office, school, household), the limit applies to all users on that network.

Rate limits are configurable by our administration team. If you believe you have been incorrectly rate-limited, contact contact@prayersoutreach.org.

Crisis Resources

If your prayer involves thoughts of suicide, self-harm, or a personal safety crisis, please also reach out to one of these resources — they are available 24/7:

  • 988 Suicide & Crisis Lifeline — call or text 988 (US)
  • Crisis Text Line — text HOME to 741741
  • International Association for Suicide Prevention — find a crisis centre near you
  • Emergency services — call 911 (US) or your local emergency number

Your submission will be reviewed by our pastoral staff, who will pray for you. But please do not wait for a response from us if you are in immediate danger.

Consequences of Violations

Rejected submissions — your prayer will not be published. Your IP address and technical data are retained for up to 30 days for abuse review, then automatically deleted.

Repeated violations — repeated submission of prohibited content may result in your IP address being blocked from the service.

Serious or illegal content — for content involving threats of violence, child exploitation, or other serious crimes, we reserve the right to preserve and report the associated technical data (IP address, device information, submission timestamp) to appropriate law enforcement authorities, including the NCMEC for child exploitation material.

Appeals

If you believe your prayer was rejected in error, you may appeal by emailing appeals@prayersoutreach.org. Please include:

  • A description or copy of the submitted prayer
  • The approximate date and time of submission
  • Why you believe it should be approved

We review all appeals and respond within 7 business days. Appeals decisions are final.

Reporting Violations

If you see content on the prayer wall that you believe violates this policy, please report it to report@prayersoutreach.org. Include a description of the content and the approximate location on the wall. We investigate all reports.

Changes to This Policy

We may update this policy to reflect new content categories, updated moderation practices, or legal requirements. The "Last updated" date at the top of this page will reflect changes.

Contact

PrayersOutreach  ·  mymiracles.org