Transparency Report
Version 1.0 · Effective 22 September 2026 · Last reviewed 22 September 2026
In short
ChatNow is not a categorised service under the Online Safety Act 2023, so no law requires us to publish a transparency report. We publish one anyway, once a year, because users, parents and regulators are entitled to know how much is reported, how quickly we act, how often we get it wrong on appeal, and how often we hand data to the authorities. This first edition covers the period from launch to 30 June 2026. Every figure comes from our moderation and complaints registers, and the definitions we used are set out in section 11.
1. Why we publish this report
Under the Online Safety Act 2023, the duty to produce transparency reports (sections 77 to 79) applies only to services placed on Ofcom's Register of Categorised Services, published on 10 July 2026. ChatNow is well below the user thresholds for Category 1, 2A or 2B and is not on that register. This report is therefore voluntary: Ofcom has not required it, has not set its content and has not verified it.
We publish it because the safety systems described in our Online Safety page and Terms of Service are only credible if their results can be seen. The report also draws on the written records we keep under section 23 of the Act. Where a number is embarrassing we publish it anyway; a transparency report that only shows good news is advertising.
2. Reporting period and scope
Reporting period: from the launch of the UK service to 30 June 2026 inclusive ("the period"). Subsequent editions will cover 1 July to 30 June.
Scope: all user-to-user functions of ChatNow available at chat-now.uk: public rooms, private messages, photo sharing, voice notes, live webcam, profiles and galleries. Figures cover guests and registered members alike. Where a metric is only meaningful for one group we say so.
Operator: Skyline Online LLC. Named person accountable for online-safety compliance: Gérald A., Head of Trust & Safety.
3. Reports received, by category
A "report" is a submission through the in-app report button on a message or profile, or an email to safety@chatnow.fr, that identifies specific content or a specific user. Categories follow the options offered in the report form: grooming and suspected under-18 users are counted under child sexual exploitation and abuse, threats to share and cyberflashing under intimate images, and stalking under threats or harassment. One report may name several messages; it is counted once.
| Category | Reports received | Upheld (action taken) | Rejected |
|---|---|---|---|
| Terrorism | [[N]] | [[N]] | [[N]] |
| Child sexual exploitation and abuse | [[N]] | [[N]] | [[N]] |
| Intimate images | [[N]] | [[N]] | [[N]] |
| Threats or harassment | [[N]] | [[N]] | [[N]] |
| Hate | [[N]] | [[N]] | [[N]] |
| Fraud | [[N]] | [[N]] | [[N]] |
| Drugs or weapons | [[N]] | [[N]] | [[N]] |
| Suicide and self-harm | [[N]] | [[N]] | [[N]] |
| Other illegal content | [[N]] | [[N]] | [[N]] |
| Breach of the Community Rules (not illegal): spam, unsolicited sexual content outside adult rooms, impersonation, abuse | [[N]] | [[N]] | [[N]] |
| Something else | [[N]] | [[N]] | [[N]] |
| Total | [[N]] | [[N]] | [[N]] |
Of the total, [[N]] reports came from people without an account (affected persons and members of the public) by email. [[N]] reports were closed as manifestly unfounded under the policy described on the Complaints & Appeals page; they are included in "Rejected".
4. Actions taken
Actions are counted by outcome, not by report: one upheld report may lead to both a removal and a ban, which are counted once each. Actions triggered by our automated systems without a user report are shown separately so that the share of proactive detection is visible.
| Action | Following a user report | Proactive (automated detection or moderator observation) | Total |
|---|---|---|---|
| Content removals (messages, images, voice notes, bios) | [[N]] | [[N]] | [[N]] |
| Temporary mutes | [[N]] | [[N]] | [[N]] |
| Temporary bans | [[N]] | [[N]] | [[N]] |
| Permanent bans (account, IP and device) | [[N]] | [[N]] | [[N]] |
| Images blocked at upload by hash-matching (Project Arachnid, StopNCII.org) | n/a | [[N]] | [[N]] |
[[N]] of the content removals in the "proactive" column were made automatically by the text classifier and logged; [[N]] of those were later reversed on appeal or on routine review (see section 6). No permanent ban was applied without human review.
5. Time to decision
Measured from the moment a report is received (timestamp of the report form or of the email arriving on our server) to the moment a decision is recorded and notified to the reporter. Median is used because a small number of complex cases would distort an average. Our published commitments are: acknowledgement within 24 hours, intimate-image reports actioned within 48 hours (section 10(3A) of the Online Safety Act 2023), standard reports decided within 7 days.
| Report type | Median time to decision | Decided within the published target |
|---|---|---|
| Intimate-image reports (48-hour route) | [[N]] hours | [[N]]% |
| Child-safety reports | [[N]] hours | [[N]]% |
| Other illegal-content reports | [[N]] hours | [[N]]% |
| Community Rules reports | [[N]] hours | [[N]]% |
6. Appeals
An appeal is a complaint by a user against a moderation decision taken against them (removal, mute, ban, reduced visibility, termination), or by a reporter against the rejection of their report. Every appeal is decided by a human reviewer who did not take the original decision, within 14 days. "Overturned" means the original decision was reversed in whole or in part and the content or account reinstated.
| Appeal type | Received | Overturned | Upheld | Median time to decision |
|---|---|---|---|---|
| Against content removal | [[N]] | [[N]] | [[N]] | [[N]] days |
| Against a mute or temporary ban | [[N]] | [[N]] | [[N]] | [[N]] days |
| Against a permanent ban or termination | [[N]] | [[N]] | [[N]] | [[N]] days |
| By a reporter against rejection of their report | [[N]] | [[N]] | [[N]] | [[N]] days |
| Complaints about ChatNow's compliance with the Online Safety Act | [[N]] | n/a | n/a | [[N]] days |
The overturn rate is the figure we watch most closely: a high rate means our first decisions are too aggressive; a rate of zero would suggest appeals are not being reviewed independently.
7. Age assurance
Every visitor, guest or member, completes a highly effective age check operated by AgeVerif before reaching any room, message, image or webcam. ChatNow receives only a pass or fail result and never sees the document, photo or estimate. Figures below are counts of results received by ChatNow; a person who tries twice is counted twice. "Failed" includes people assessed as under 18 and people who abandoned the check before a result.
| Metric | Number |
|---|---|
| Age checks started | [[N]] |
| Age checks completed with a pass | [[N]] |
| Age checks failed or abandoned | [[N]] |
| Complaints from users who say they were wrongly assessed as under 18 | [[N]] |
| Accounts closed because the user was found to be under 18 after passing the check | [[N]] |
We do not receive, and therefore cannot publish, a breakdown by method (facial estimation, document, credit card, digital wallet): that information stays with the provider under the double-anonymity design described in the Age Assurance Policy.
8. Child sexual exploitation and abuse: reports to the NCA
Since 7 April 2026, the Online Safety (CSEA Content Reporting by Regulated User-to-User Service Providers) Regulations 2026 require us to report child sexual exploitation and abuse content detected on the service to the National Crime Agency through its CSEA Industry Reporting Portal. Detection comes from Project Arachnid hash-matching of every uploaded image, from the text classifier, from moderators and from user reports.
| Metric | Number |
|---|---|
| Reports made to the NCA | [[N]] |
| Of which triggered by a Project Arachnid hash match | [[N]] |
| Of which triggered by a user report or moderator observation (including grooming) | [[N]] |
| Accounts permanently banned in connection with these reports | [[N]] |
| URLs or images also referred to the Internet Watch Foundation | [[N]] |
We publish counts only. We do not publish any detail that could identify a victim, a suspect or an ongoing investigation, and where the NCA or a police force asks us to, because publication would prejudice an investigation, we defer publication of a figure to the next edition and say so.
9. Requests from law enforcement and authorities
Our Law Enforcement & Authorities Guidelines describe what data we hold, for how long, and the legal basis we require before disclosing it. A "request" is a formal written demand received at legal@chatnow.fr from a UK police force, the NCA, a court, Ofcom, the ICO or another public authority with statutory powers. "Complied with" means at least some of the requested data was disclosed; partial disclosures count as complied with. Requests refused for lack of legal basis, or where we held no data, are counted under "Not complied with".
| Type of request | Received | Complied with | Not complied with | Accounts affected |
|---|---|---|---|---|
| Court orders (including Norwich Pharmacal orders) | [[N]] | [[N]] | [[N]] | [[N]] |
| Police and NCA data requests (other than court orders) | [[N]] | [[N]] | [[N]] | [[N]] |
| Emergency disclosure requests (imminent risk to life) | [[N]] | [[N]] | [[N]] | [[N]] |
| Preservation requests | [[N]] | [[N]] | [[N]] | [[N]] |
| Information notices or correspondence from Ofcom or the ICO | [[N]] | [[N]] | [[N]] | n/a |
| Requests from authorities outside the UK | [[N]] | [[N]] | [[N]] | [[N]] |
10. Copyright and defamation notices
Copyright notices follow the procedure on our Copyright & Intellectual Property Complaints page; defamation notices of complaint follow the procedure on our Defamation Complaints page (section 5 of the Defamation Act 2013 and the Defamation (Operators of Websites) Regulations 2013).
| Notice type | Valid notices received | Content removed | Counter-notices or poster responses received | Content reinstated |
|---|---|---|---|---|
| Copyright and trade-mark notices | [[N]] | [[N]] | [[N]] | [[N]] |
| Defamation notices of complaint | [[N]] | [[N]] | [[N]] | [[N]] |
Notices that did not meet the statutory or procedural requirements (for example a defamation notice without the complainant's name or without an explanation of why the statement is defamatory) are not counted as valid; [[N]] such notices were returned with an explanation of what was missing. [[N]] accounts were terminated under the repeat-infringer policy.
11. Methodology and definitions
11.1 Sources
All figures are extracted from three internal systems: the moderation register (every report, automated flag, decision and sanction, with timestamps and the identity of the reviewer), the complaints and appeals register, and the legal-requests log kept at legal@chatnow.fr. Age-assurance counts come from the pass/fail results returned to our servers by AgeVerif. No figure is estimated or sampled; each is a count over the whole period. Figures were extracted on a single date after the end of the period and were reconciled by a second member of staff.
11.2 Counting rules
- Reports are counted when received, and their outcome is counted in the period in which the decision was made. A report received on 29 June and decided on 2 July appears in "received" this year and in "upheld/rejected" next year.
- Where a reporter chose a category that did not match the content, the report is counted under the category the reviewer assigned.
- Guests and members are counted the same way. Guest content is auto-deleted 24 hours after the guest leaves; the moderation record of a decision about it is kept for 12 months and is what these figures draw on.
- A ban "by account, IP and device" is one ban. Later attempts by the same person to return, blocked automatically by the ban list, are not counted as new bans.
- Median time is the middle value when all durations are sorted; where the count is even, the mean of the two middle values.
- Numbers below 5 in a category that could identify an individual (for example, a single NCA report from a small city) are published as "fewer than 5".
11.3 Limitations
This report counts what we saw. It cannot count illegal content that was never reported, never flagged and never seen by a moderator, and it should not be read as a measure of how much harmful content exists on the service. Report volumes rise when a reporting tool is easier to find, so an increase between editions is not by itself evidence that the service became less safe. The classifier's reversal rate in section 4 is our best available proxy for automated over-removal, but it only captures errors that someone appealed or that routine review caught. The figures have not been audited by a third party.
12. Next publication and archive
The next edition will cover 1 July 2026 to 30 June 2027 and will be published within three months of the end of that period, that is by 30 September 2027, at this address. Each edition is published on the same schedule thereafter.
Historical editions are kept: when a new edition is published, the previous one remains available from a link at the top of this page and is listed in the Legal Documents Version History. We do not silently revise a published figure; if an error is found after publication, the correction is made on the page with a dated note explaining what changed and why.
Questions about this report can be sent to safety@chatnow.fr. Requests from regulators or researchers for additional breakdowns are considered case by case, subject to the privacy of the people involved.