Apple is killing Rave But we are fighting back.

What Apple Did

After ten years and 73 million iOS downloads, Apple shut Rave down.

Removed from the App Store

Apple removed Rave from the App Store without notice, blocking new downloads, critical updates, and reinstalls.

Blocked on Mac

Apple revoked Rave's developer certificate, causing macOS to block Rave with a false malware warning.

11.4 million users locked out

Apple disabled Sign in with Apple for Rave, locking users out of accounts they'd used for years.

Silenced notifications

Rave iOS users stopped receiving invitations and messages from their friends.

How we tried to fix it

For months, we tried to work with Apple. They refused.

  1. We asked what we did wrong

    Apple cited a single, vague clause of their developer agreement and refused to explain further.

  2. Their reasons kept changing

    We kept asking for clarification, but Apple's reasons for removing Rave kept changing.

  3. They asked for time to "investigate"

    Months after removing Rave, they said they needed to "investigate" why they took us down.

  4. They refused to negotiate, and eventually stopped responding

    Apple's final message: "We consider this matter closed."

Taking the fight global

Rave has users around the world. We're fighting for our right to exist.

Setting the record straight

  • Apple says Rave lacked content moderation.

    That's not true. The tables below compare Rave's industry-leading child protection technologies with Apple's own (lack of) content moderation.

Rave vs Apple's own products

Rave iCloud iMessage SharePlay Books Safari
CSAM Hash Matching Yes. Cloudflare CSAM Scanning checks NCMEC/IWF hashes as Rave's known-material layer. No. Built, delayed, killed. Apple announced iCloud Photos CSAM matching in 2021, paused the rollout after backlash, then killed NeuralHash in 2022 without launching an iCloud Photos replacement. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash, and West Virginia alleges Apple knowingly left iCloud without effective CSAM detection. No. Built, delayed, killed. Apple announced iCloud Photos CSAM matching in 2021, paused the rollout after backlash, then killed NeuralHash in 2022 without launching an iCloud Photos replacement. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash, and West Virginia alleges Apple knowingly left iCloud without effective CSAM detection. No. Built, delayed, killed. Apple announced iCloud Photos CSAM matching in 2021, paused the rollout after backlash, then killed NeuralHash in 2022 without launching an iCloud Photos replacement. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash, and West Virginia alleges Apple knowingly left iCloud without effective CSAM detection. No. Built, delayed, killed. Apple announced iCloud Photos CSAM matching in 2021, paused the rollout after backlash, then killed NeuralHash in 2022 without launching an iCloud Photos replacement. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash, and West Virginia alleges Apple knowingly left iCloud without effective CSAM detection. No. Built, delayed, killed. Apple announced iCloud Photos CSAM matching in 2021, paused the rollout after backlash, then killed NeuralHash in 2022 without launching an iCloud Photos replacement. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash, and West Virginia alleges Apple knowingly left iCloud without effective CSAM detection.
AI CSAM Detector Yes. CSAM-detecting AI. Rave developed an AI detector for unknown CSAM, pornography, bestiality, and gore. State of the art validation accuracy exceeding 99.92% for detecting CSAM. Live demo available at a-eye.com. No. No public novel-CSAM classifier. Apple’s public system is Communication Safety: nudity warnings for child-accounts, not CSAM/grooming classification. Australia’s eSafety reported Apple did not proactively detect new CSEA material in iCloud Photos and had no measures for live CSEA on video calls such as FaceTime; Apple publishes no visual-abuse model metrics. No. No public novel-CSAM classifier. Apple’s public system is Communication Safety: nudity warnings for child-accounts, not CSAM/grooming classification. Australia’s eSafety reported Apple did not proactively detect new CSEA material in iCloud Photos and had no measures for live CSEA on video calls such as FaceTime; Apple publishes no visual-abuse model metrics. No. No public novel-CSAM classifier. Apple’s public system is Communication Safety: nudity warnings for child-accounts, not CSAM/grooming classification. Australia’s eSafety reported Apple did not proactively detect new CSEA material in iCloud Photos and had no measures for live CSEA on video calls such as FaceTime; Apple publishes no visual-abuse model metrics. No. No public novel-CSAM classifier. Apple’s public system is Communication Safety: nudity warnings for child-accounts, not CSAM/grooming classification. Australia’s eSafety reported Apple did not proactively detect new CSEA material in iCloud Photos and had no measures for live CSEA on video calls such as FaceTime; Apple publishes no visual-abuse model metrics. No. No public novel-CSAM classifier. Apple’s public system is Communication Safety: nudity warnings for child-accounts, not CSAM/grooming classification. Australia’s eSafety reported Apple did not proactively detect new CSEA material in iCloud Photos and had no measures for live CSEA on video calls such as FaceTime; Apple publishes no visual-abuse model metrics.
Child Predator Detector Yes. “Automod” AI system. Rave scans all public and private chats with AI; automatic escalations to full account reviews trigger permanent account/device bans, which have been issued for 120,000 users. No. No pedophilia or grooming scan. iCloud Photos stores and syncs libraries; Apple publishes no grooming or predatory-context review for photo libraries. No. No pedophilia or grooming protection. Conversations are not monitored. Communication Safety is on by default for child accounts, and only checks photos/videos for nudity; it is not conversation-level grooming detection. No. No live-call predator detection. FaceTime calls/video messages can trigger child nudity warnings for child accounts; Apple publishes no live-audio or SharePlay context moderation. No. Storefront content, not safety screening. Apple Books features explicit authority/minor-adjacent erotica, such as teacher/student, babysitter, and stepdaughter, themes commonly associated with child predators. No. No browsing-level predator detection. Safari restrictions are configurable filters, not grooming/pedophilia detection over browsing behavior.
Age Verification Yes. Mandatory 18+ facial age verification. AI age estimation is required before mature content can be viewed. No. Mostly self-declared Apple age. Apple Accounts still begin with user-entered birthdays; adult confirmation applies only to limited countries/actions. Apple does turn on child-account protections such as Communication Safety and web content limits, but its Declared Age Range API is a consent/regulatory signal for apps - not a universal facial/ID gate before iCloud, iMessage, FaceTime/SharePlay, Books, or Safari. No. Mostly self-declared Apple age. Apple Accounts still begin with user-entered birthdays; adult confirmation applies only to limited countries/actions. Apple does turn on child-account protections such as Communication Safety and web content limits, but its Declared Age Range API is a consent/regulatory signal for apps - not a universal facial/ID gate before iCloud, iMessage, FaceTime/SharePlay, Books, or Safari. No. Mostly self-declared Apple age. Apple Accounts still begin with user-entered birthdays; adult confirmation applies only to limited countries/actions. Apple does turn on child-account protections such as Communication Safety and web content limits, but its Declared Age Range API is a consent/regulatory signal for apps - not a universal facial/ID gate before iCloud, iMessage, FaceTime/SharePlay, Books, or Safari. No. Mostly self-declared Apple age. Apple Accounts still begin with user-entered birthdays; adult confirmation applies only to limited countries/actions. Apple does turn on child-account protections such as Communication Safety and web content limits, but its Declared Age Range API is a consent/regulatory signal for apps - not a universal facial/ID gate before iCloud, iMessage, FaceTime/SharePlay, Books, or Safari. No. Mostly self-declared Apple age. Apple Accounts still begin with user-entered birthdays; adult confirmation applies only to limited countries/actions. Apple does turn on child-account protections such as Communication Safety and web content limits, but its Declared Age Range API is a consent/regulatory signal for apps - not a universal facial/ID gate before iCloud, iMessage, FaceTime/SharePlay, Books, or Safari.
Pornography Controls Yes. Off by default; third-party only. Mature co-viewing is locked until verified 18+. Rave does not host co-viewed streams; it age-gates incidental third-party mature content under the same UGC framework Apple applies to apps like X and Reddit. No. User libraries, no adult gate. iCloud Photos has no restrictions on storing and sharing pornography. Partial. Adult accounts unrestricted. Communication Safety warns children about nude photos/videos; adult iMessage accounts have no mature-content gate. No. No SharePlay age gate. SharePlay/FaceTime can be used to co-view or screen-share adult web content without verified 18+ access control. No. Apple sells erotica. Apple’s Books rules prohibit explicit content, Apple Books defaults to “explicit” rather than “clean” and features pornographic books, audiobooks, and photobooks, not fully hidden even when set to “clean”. No. No restrictions by default. Web content limits are on by default for child accounts (age varies by region), which prevent access to some but not all pornographic websites.
Lawsuits & Investigations Yes. Apple’s Review in February 2025 found no CSAM and praised Rave’s moderation tech; 6 months later, Apple removed Rave for “Dishonest or Fraudulent Activity”; two months later Apple cites “a user complaint alleging” CSAM, but never claimed to have found CSAM on Rave. No. Apple's own services are under intense CSAM scrutiny. West Virginia's AG complaint quotes Apple calling itself the “greatest platform for distributing child porn” and saying it had “chosen to not know” enough to act. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash without a replacement; AMY v. Apple brings claims by CSAM victims; the NSPCC accused Apple of underreporting; and eSafety found gaps on new CSEA material and live CSEA on video calls. No. Apple's own services are under intense CSAM scrutiny. West Virginia's AG complaint quotes Apple calling itself the “greatest platform for distributing child porn” and saying it had “chosen to not know” enough to act. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash without a replacement; AMY v. Apple brings claims by CSAM victims; the NSPCC accused Apple of underreporting; and eSafety found gaps on new CSEA material and live CSEA on video calls. No. Apple's own services are under intense CSAM scrutiny. West Virginia's AG complaint quotes Apple calling itself the “greatest platform for distributing child porn” and saying it had “chosen to not know” enough to act. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash without a replacement; AMY v. Apple brings claims by CSAM victims; the NSPCC accused Apple of underreporting; and eSafety found gaps on new CSEA material and live CSEA on video calls. No. Apple's own services are under intense CSAM scrutiny. West Virginia's AG complaint quotes Apple calling itself the “greatest platform for distributing child porn” and saying it had “chosen to not know” enough to act. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash without a replacement; AMY v. Apple brings claims by CSAM victims; the NSPCC accused Apple of underreporting; and eSafety found gaps on new CSEA material and live CSEA on video calls. No. Apple's own services are under intense CSAM scrutiny. West Virginia's AG complaint quotes Apple calling itself the “greatest platform for distributing child porn” and saying it had “chosen to not know” enough to act. The Kansas AG demanded iCloud safeguards after Apple abandoned NeuralHash without a replacement; AMY v. Apple brings claims by CSAM victims; the NSPCC accused Apple of underreporting; and eSafety found gaps on new CSEA material and live CSEA on video calls.

Rave vs apps still on the App Store

Rave Kik Telegram Roblox BIGO Live Twitter Discord Reddit Snapchat Tumblr WhatsApp Signal
CSAM Detector Yes. Hashes + novel visual AI. Cloudflare hashes known CSAM; Rave Safety/a-eye.com add novel visual detection with validation accuracy of 99.92% for detecting CSAM. No. No public stack. Child-safety advocates document severe failures; no public hash/novel-model transparency. Partial. Public surfaces only. Telegram joined IWF for public CSAM hashes; private and Secret Chats remain unscanned. Partial. Uploaded assets, unclear stack. Roblox says uploaded images are evaluated for CSAM and reported to NCMEC, and its transparency reports disclose enforcement. Partial. Claimed AI/partnerships. BIGO claims AI moderation, but the NYT documented exploitation getting through. Partial. PhotoDNA history, weakened trust layer. Twitter adopted PhotoDNA, but Thorn cut ties and NBC reported CSAM-trading bots. Yes. Hashing + novel model. Discord describes CSAM matching and a novel-CSAM model, but publishes no F1/recall metrics. Yes. Hashing/reporting. Reddit describes PhotoDNA/CSAI tools; no public novel classifier or grooming model. Partial. Tools, but limits. Snap says it uses CSAM detectors, but eSafety describes their use mainly on public surfaces and reported private media, while New Mexico alleges CSAM-database and response gaps. Partial. Historic PhotoDNA, limited current detail. Tumblr was removed over CSAM in 2018; current novel-model detail is sparse. Partial. Not scanning encrypted messages. Safety tools rely on reports; encryption means private message content is not scanned, and reporting has documented CSAM trading on WhatsApp. No. No content scanning. Signal says messages and calls cannot be accessed by Signal or third parties; its president says mandated scanning undermines encryption.
Child Predator Detector Yes. “Automod” AI system. All public/private text is reviewed in context; escalations trigger account-level review and device bans. No. Predator-risk design. NCOSE describes Kik as a predator haven; no public grooming detector. Partial. Public spaces only. AI moderation excludes private and Secret Chats. Partial. Sentinel detects harmful text patterns, but the 2026 audit found it can be evaded by child predators. Partial. Claims, but failures reported. BIGO AI claims coexist with NYT findings of livestreamed exploitation. No. No public detector. Safety-team cuts and no disclosed grooming classifier. Partial. Teen assist only. Teen Safety Assist exists; Discord says adult DMs are not proactively read by default. Partial. Reports/mods, no public classifier. Reddit CSE tools rely on tools, reports, admins, and communities. Partial. Warnings, not full detection. Stranger warnings are limited; New Mexico alleges sextortion reports went unaddressed. Partial. Labels, not predator detection. Community Labels filter sexual content but are not grooming detection. Partial. Metadata/user reports only. WhatsApp can act on reports and some account signals, but encrypted messages prevent monitoring. No. None by design. Signal’s design prevents content access, so there is no conversation-level grooming scan.
Age Verification Yes. Mandatory 18+ facial gate. AI age estimation is required before accessing mature content. Partial. UK-only proof. Kik says UK users verify; elsewhere it is mostly policy age, not proof. No. Policy age only. Telegram ToS and phone numbers are not universal age proof. Yes. Real checks for chat features. Roblox requires age checks for communication features. Partial. Claimed/limited. BIGO claims facial recognition, but no universal public proof gate is shown. Partial. Legal-triggered. X verifies/estimates age where required, not as a universal mature-content gate. Partial. Rolling out. Discord age assurance is partial/age-restricted, not universal. Partial. Regional/NSFW contexts. Reddit asks for age in affected contexts; not global proof for all users. Partial. Prompted/contextual. Snap verifies if prompted depending on jurisdiction/service availability. No. No universal proof. Community Labels and DOB are not facial/ID verification. No. Policy age only. WhatsApp teen guidance and account rules are not facial or ID age verification. No. Policy age only. Signal’s terms do not create an in-app facial/ID age gate.
Pornography Controls Yes. Off until verified. Mature content is third-party only and inaccessible until the 18+ facial gate is passed. No. Sexual-content risk. NCOSE/Bark reported severe sexual-content concerns. No. Legal adult content persists. Telegram ToS bans illegal abuse content, not all adult content. Partial. Prohibited, but exposed. Hindenburg alleged adult/CSAM trading groups. No. Prohibited, but exposed. NYT documented livestreamed child exploitation. Partial. Adult content allowed. X policy change formally permits adult content. Partial. NSFW allowed. Discord filters manage sensitive media and age-gated spaces. Partial. Massive NSFW access. Apple permits Reddit despite large porn access under UGC rules. Partial. Prohibited, private risk. New Mexico complaint alleges serious exploitation gaps. Partial. Nudity allowed. Tumblr reversed its nudity ban; explicit sex acts remain barred. Partial. Public/business only. WhatsApp can receive user reports, but encryption prevents scanning of messages/groups. No. Unenforceable by design. Signal cannot access message content and opposes mass scanning.
Lawsuits & Investigations Yes. Apple investigated and found no CSAM, and approved of Rave's content moderation system. No. Repeated child-safety cases. Doe v. Kik plus federal Kik CSAM prosecutions. No. CEO charged, regulators probing. Durov indictment tracker, Ofcom probe, and eSafety fine. No. Major litigation. AG suits, federal MDL, settlements, and independent audit findings. No. Exploitation investigation. NYT investigation led to a brief App Store takedown. No. Child-abuse lawsuit survived key claims. Doe v. Twitter/X revived claims; Thorn cut ties. No. State AG + Senate scrutiny. New Jersey sued Discord; Blackburn and others scrutinized NCMEC reporting. Partial. CSAM removals reported. Reddit reported 31,574 CSAM removals and 40,243 NCMEC reports in H2 2022. No. AG suit plus safety litigation. New Mexico sued Snap claiming “Snap Knowingly Contributed to Harm Amongst Children” Partial. Prior iOS removal. Tumblr was removed in 2018 for 27 days over CSAM and restored after an adult-content ban, which it later reversed. No. WhatsApp-specific CSAM pressure. India demanded action on CSAM; Wired described how encryption hides CSAM, and eSafety put WhatsApp on notice. Partial. Regulatory pressure. Signal would rather leave than weaken encryption.
App Store Status No. Apple removed Rave for “Dishonest or Fraudulent Activity”, later cites “a user complaint alleging” CSAM. Yes. Still listed. Kik remains on the App Store despite its predator/CSAM record. Yes. Still listed. Telegram remains on the App Store despite Durov charges and an eSafety fine. Yes. Still listed. Roblox remains on the App Store, and is heavily promoted by Apple. Yes. Restored/listed. Allowed back on the App Store after promising to improve moderation. Yes. Still listed. X remains on the App Store despite Thorn cutoff and Doe litigation. Yes. Still listed. Discord remains on the App Store despite NJ AG suit and NCMEC-reporting scrutiny. Yes. Still listed. Reddit remains on the App Store despite massive porn access. Yes. Still listed. Snapchat remains on the App Store despite NM AG litigation and fentanyl litigation. Yes. Restored/listed. Tumblr remains on the App Store after CSAM removal and return of adult-content. Yes. Still listed. WhatsApp remains on the App Store despite CSAM reports, eSafety scrutiny, and encryption preventing scanning. Yes. Still listed. Signal remains on the App Store despite its no-scanning stance.

Press release

Our full statement for journalists covering this story.