Hash matching plus forensic classifiers, correlated with grooming and sextortion signals. One API for image, video and document scanning. EU-resident, zero retention.
What Tuteliq detects
Full pipeline documented in the API reference, with grooming and sextortion detectors in /grooming and /sextortion.
Known CSAM (hash-based)
Perceptual and cryptographic hash matching against trusted hash lists to flag re-uploads of previously reported material.
Novel and AI-generated CSAM
Forensic classifiers designed to flag suspected previously unidentified and synthetic CSAM, with behavioural signals for case context. Detection is not guaranteed.
Multi-modal coverage
Runs across images, video frames and documents. Signals combine with grooming, sextortion and coercion detection for full case context.
Auditable case files
Every detection produces a signed case record with model version, confidence, jurisdiction and reviewer trail, ready for NCMEC, IWF or eSafety referral.
Novel CSAM: beyond a known-material hash match
Novel CSAM means previously unidentified child sexual abuse material with no matching reference in the hash list being used. It is not a synonym for AI-generated CSAM: previously unidentified camera-recorded material can also be novel.
Hash matching compares a file's fingerprint with known references. Classification evaluates the content for suspected abuse without requiring that exact reference match. These approaches complement each other; neither guarantees complete detection.
Tuteliq's forensic and synthetic-content analysis is designed to flag suspected material for your team's response. Combine media signals with grooming and sextortion context where available. Test false positives and missed detections on lawful evaluation material, set escalation rules and preserve human review. Do not upload or share illegal material to try a detector.
Hash matching vs classifiers vs Tuteliq
Hash matching needs known references. Classifiers can flag suspected new material. Tuteliq adds behavioural context when your platform supplies it.
Hash matching
Generic classifier
Tuteliq
Known CSAM (re-uploads)
Detects
Novel CSAM (no reference match)
Needs a reference hash
Can flag suspected material
Designed to flag suspected material
AI-generated / synthetic CSAM
Only if a reference matches
Depends on the model
Synthetic-content analysis
Grooming context around the media
No
Yes
EU-resident processing
Varies
Zero raw-content retention
Where teams deploy CSAM detection
Social & UGC
Scan uploads and DMs the moment media hits your service, before it fans out to other users.
Cloud storage & backup
Detect on write and on share-link creation without keeping copies of the flagged content.
Dating & companion apps
Combine CSAM detection with grooming and sextortion signals to escalate the full pattern, not just the file.
Marketplaces & payments
Screen listing images and profile media to block CSAM-adjacent monetisation.
Safeguards and compliance
CSAM detection carries higher duty of care than any other moderation task. Tuteliq's approach is documented in the Trust Center, compliance overview and DPA.
- Signed, tamper-evident case files
- Human review workflows
- NCMEC / IWF / eSafety-ready records
- DSA Article 18 traceability
Frequently asked questions
Stand up CSAM detection this week
EU-resident, zero-retention CSAM and synthetic content detection with a signed audit trail.