Content Moderation Policy
How we detect, review and act on behaviour that breaks our rules - and how you challenge a decision you think is wrong.
Last updated 23 September 2026
At a glance
The full policy below is what applies.
An automated system analyses every session in real time and ends it immediately where it detects a likely violation.
No account is suspended or removed on an automated flag alone.
Minor breaches earn a warning and accumulate. Severe ones end an account the first time.
We notify you of any action against your account, explain why and give you a route to appeal.
Talking about self-harm never results in action against you. It triggers support, not enforcement.
Contents
1Scope
This policy governs how we moderate conduct on Miku. It applies to every interaction on the platform - sessions, profiles and reports.
It describes enforcement. What is actually prohibited is set out in our Community Guidelines and the two documents should be read together. Our Terms of Use, Privacy Policy and Compliance Statement also apply.
By using the app you agree to be moderated in the way described here.
2How moderation works
Three things bring a potential violation to our attention. Most cases come through the first two.
Layer one
Automated detection
Real-time analysis of every session. Ends the session on detection, without waiting for anyone to review it.
Layer two
Reports from users
Anyone can report during a session or after.
Layer three
Proactive review
Patterns we watch for ourselves - repeated flags, unusual session behaviour, clusters of complaints about one account.
Detection is only the start. Ending a session is automatic; acting against an account is not. Every enforcement decision involving suspension or removal is made by a person.
3Automated detection
Session audio and video are analysed in real time by an automated safety system operated by a specialist provider on our behalf.
What it looks for
- Nudity and sexual content
- Sexual language and solicitation
- Abuse, threats and harassment
- Requests for money, one-time passwords or payment details
- Attempts to exchange phone numbers, social handles or other contact details
- Attempts to move the conversation off the platform or to arrange an offline meeting
What happens on detection
- The session ends immediately. Neither party has to do anything.
- Billing stops at the moment the session ends. You are not charged for time you did not use.
- The flag and the material that triggered it, is preserved for review.
- The case enters the review queue described in section 4.
A flag is a signal, not a finding. It does not by itself mean you have done something wrong and it does not by itself result in a penalty.
4Human review
Cases are reviewed by a trained safety team.
- Reviewers are trained on our Community Guidelines, on how to grade severity and on handling sensitive material.
- Access to session material is restricted to reviewers handling a specific case.
- Every access is logged, including who opened the material and why.
- Reviewers are bound by confidentiality obligations in their contracts.
- Decisions that would permanently remove an account are confirmed by a second reviewer.
Reviewers work to the same severity framework for everyone. The tenure or spending history of an account does not change the outcome.
5Reports from users
You can report from the menu during a session or from your session history afterwards. Reporting during a session does not require you to end it first.
Our safety team responds within 24 hours.
Reports are confidential. The person reported is never told who reported them.
You can block someone whether or not you report them. Blocking is immediate and needs no reason.
Reports made to harass someone or repeatedly and without basis, are themselves a violation of our Community Guidelines and are treated accordingly. This does not apply to a report made in good faith that turns out to be mistaken - that is welcome and we would far rather look at something that comes to nothing.
6Severity and enforcement
Every confirmed violation is graded into one of four tiers. The tier determines what happens.
| Tier | Examples | Action |
|---|---|---|
| 1 - Minor |
| Warning. Content removed where relevant. Counts as a strike. |
| 2 - Serious |
| Session ended. Warning or suspension on repetition. Counts as a strike. |
| 3 - Severe |
| Session ended. Suspension on a first instance, permanent removal on a second. Balance forfeited where our Terms provide. No refund. |
| 4 - Critical |
| Immediate permanent removal, first instance, regardless of history. Referred to law enforcement. Evidence preserved. |
Within a tier, we weigh whether the conduct was deliberate, whether it was repeated, the effect on the person affected and whether the account has prior violations. Context can move a case up or down a tier, but it cannot move anything out of Tier 4.
Strikes from Tier 1 and Tier 2 violations accumulate. Three strikes within six months results in permanent removal. Strikes expire after six months where no further violation occurs.
7Immediate removal
Certain conduct ends an account on the first occasion, whatever the account's history and regardless of any explanation offered:
- Anything that sexualises or endangers a child
- Trafficking, exploitation or coercion
- Credible threats of serious violence
- Recording, distributing or threatening to distribute another user's session material
- Extortion
In these cases we preserve the evidence, refer the matter to the appropriate authorities and cooperate with any resulting investigation.
8Notice and your right to dispute
Where we remove content or act against your account, we tell you:
- What action we have taken
- Which part of our Community Guidelines it relates to
- When it takes effect and, for a suspension, when it ends
- How to dispute it
Notice is given in the app and by SMS or email. We may act before giving notice where there is a risk to someone's safety, a legal obligation or a Tier 4 violation - but you are told afterwards either way, unless telling you would prejudice an investigation or the law prevents it.
9Appeals
If you think we have got it wrong, tell us.
- Write to info@mikuapp.live or to our Grievance Officer, quoting your registered mobile number.
- A reviewer who was not involved in the original decision looks at the case again.
- Where an automated system contributed to the decision, you may specifically request human review.
- We acknowledge your appeal within 24 hours.
If the appeal succeeds, the action is reversed, any strike is removed and a suspension is lifted. Where a balance was forfeited in error it is restored - see the Charges and Refund policy.
If you remain dissatisfied, you may escalate to the Grievance Appellate Committee constituted under Rule 3A of the IT Rules, within thirty days of our decision.
10Evading enforcement
Creating a new account to get around a suspension or removal is itself a violation. Where we identify evasion, we remove the new account and may extend the original penalty.
We use account, device and payment signals to identify evasion. We do not use them to link accounts for any other purpose.
11Distress and self-harm
People come to this platform to talk about hard things. Some of those conversations touch on self-harm, suicidal thoughts or acute distress. This is the reason the platform exists and we handle it as a support situation, never as an enforcement one.
Talking about your own distress, including thoughts of self-harm, never results in a strike, a suspension or any action against your account. Our Community Guidelines prohibit encouraging or glorifying self-harm and describing your own experience is not that.
What the other person does
- Says plainly that they are not trained for this, rather than attempting to counsel.
- Shares the crisis resources below and offers to stay on the line while the person makes contact.
- Encourages the person to involve a family member or friend.
- Stays on the call for as long as the person wants them to.
- Flags the session to our safety team afterwards, so the person can be followed up with support.
What our team does
- Treats the flag as a priority case, handled ahead of the ordinary queue.
- Reaches out to the person with resources, without any suggestion of enforcement.
- Contacts emergency services where there is reason to believe someone is in immediate danger.
- Records the case separately from the enforcement log, so it never contributes to an account's violation history.
Tele-MANAS: 14416 - free, confidential, 24 hours a day across India. Emergency services: 112.
12Child safety escalation
Any indication that a child is involved - as a user, in shared material or as the subject of a conversation - is escalated immediately to a senior member of the safety team and is not handled through the ordinary queue.
- The account is suspended pending review, not after it.
- Evidence is preserved and is not deleted on the normal retention schedule.
- The matter is referred to the appropriate authorities.
- Where an account is found to belong to someone under 18, it is closed and its data deleted and the parent or guardian contact route in our Privacy Policy applies.
We do not weigh context, intent or account history in this category.
13Moderation of the people you talk to
The people you talk to are held to the Community Guidelines in full. In several respects the standard applied to them is higher, because the power in a session is not evenly distributed.
- Asking a user for contact details, money or an offline meeting is treated as Tier 3 at minimum.
- Repeated low-rated sessions or a cluster of reports triggers proactive review even where no single report is conclusive.
- They may not record, transcribe or screenshot a session under any circumstances.
They complete KYC verification. Verification establishes identity. It is not a prediction of conduct and it does not reduce the standard they are held to afterwards.
14Recording is prohibited
No user may record, transcribe, screenshot or otherwise capture any part of a session - audio, video or chat - by any means, including a second device.
Recording another person's session is a Tier 3 violation. Distributing a recording, or threatening to, is Tier 4 and is referred to law enforcement.
If you believe someone has recorded you, report it immediately. Tell us what you saw or heard that made you think so and when in the session it happened.
15What moderation cannot do
We would rather be straight about the limits than imply a guarantee.
- Automated detection is imperfect. It misses things and it ends sessions that turn out to be fine. If a session of yours was ended in error, tell us and we will look at it and refund the affected time.
- We act after the fact, mostly. Ending a session stops something continuing; it does not undo what was already said.
- Verification establishes identity, not character. KYC tells us who someone is. It does not tell us how they will behave.
- We cannot moderate what leaves the platform. This is the reason our rules on contact details and offline meetings are as strict as they are.
Your own judgement remains the most important safeguard.
16Contact
Report anything that breaks our rules from inside the app or write to info@mikuapp.live. Our safety team responds within 24 hours.
We acknowledge grievances within 24 hours.
17Review of this policy
We review this policy at least once a year and whenever our practices or the law change materially. Where a change affects how enforcement works, we will tell you in the app before it takes effect.
Please check this page from time to time so that you know what the current version says.