analyze_image

Quick comprehensive safety check (bullying + unsafe)

Analyze PDF documents for safety concerns — per-page multi-endpoint detection with chain-of-custody hashing (max 50MB, 100 pages)

Analyze emotional content and mental state indicators — accepts single text or full conversations

Analyze video files for safety concerns via key frame extraction (supports mp4, mov, avi, webm, mkv)

Transcribe audio and run safety analysis on the transcript

Analyze up to 50 items in a single request — all twelve detection types (bullying, grooming, unsafe, emotions, socialengineering, appfraud, romancescam, mulerecruitment, gamblingharm, coercivecontrol, vulnerabilityexploitation, radicalisation) — ideal for bulk triage

Apply one decision across many incidents in a single call

Cancel an active session (no credits consumed)

Create a session for age or identity verification — returns a URL for the user to complete the flow

Create a new webhook endpoint

Detect app-based fraud (fake investment platforms, phishing apps, subscription traps)

Analyze text for bullying, harassment, and gaming toxicity — including coded slang, emoji, and deliberate filter evasion, with context that tells trash-talk apart from genuine harm

Detect coercive control patterns (isolation, financial control, monitoring, threats)

Detect gambling-related harm indicators (chasing losses, concealment, distress)

Detect grooming patterns and predatory behavior in conversations

Detect money mule recruitment tactics (easy-money offers, bank account sharing)

Detect radicalisation indicators (extremist rhetoric, us-vs-them framing, ideological grooming)

Detect romance scam patterns (love-bombing, financial requests, identity deception)

Detect social engineering tactics (pretexting, urgency fabrication, authority impersonation)

Dual-signal forensics: transcript + mel spectrogram vision + quantitative audio statistics

6-signal forensic pipeline: vision AI, EXIF metadata, pixel stats, C2PA Content Credentials, watermarks, pHash

Detect AI-generated text across 10 child-safety categories (synthetic CSAM, deepfake scripts, AI grooming)

5-track analysis: per-frame vision, temporal face consistency, lip-sync correlation, spectral audio, transcript

Identify unsafe content (self-harm, violence, explicit material)

Detect exploitation of vulnerable individuals (elderly, disabled, financially distressed)

Create incident reports from conversations

Generate age-appropriate guidance for safety situations

Query the audit log

Fetch the signed receipt for a past inference

Full detail for one incident, including the risk trajectory across messages

Incident volume bucketed by hour, day or week, split by severity

Counts by category, severity, source, status and platform over a window

Poll session status — returns full document intelligence (MRZ, barcode, authenticity, face match, liveness)

Paginated, filterable incident list

List all configured webhooks

Moderator triage console: the unreviewed queue, the next item, and — optionally — your own analysis trace and recommended decision, rendered for human sign-off. Read-only

Record a moderator decision (confirm / downgrade / escalate / reclassify / dismiss) with a signed receipt

Update webhook configuration