Skip to content

CHANGELOG.md

Latest commit

 

History

History
72 lines (54 loc) · 3.23 KB

CHANGELOG.md

File metadata and controls

72 lines (54 loc) · 3.23 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[2.5.0] - 2026-03-15

Added

  • country context field — Pass ISO 3166-1 alpha-2 country code (e.g., "GB", "US", "SE") in the context object to receive geo-localised crisis helpline data in detection responses. Falls back to user profile country if omitted.

Improved

  • Action escalation for minors — All detection endpoints now enforce minimum flag_for_review when harm is detected and the subject is a minor. Criminal indicators (sextortion, trafficking, CSAM) targeting minors automatically escalate to immediate_intervention.
  • Risk score distribution — Detection prompts now instruct the LLM to use graduated scoring across the full 0.0–1.0 range instead of clustering around a single value.
  • Evidence tactic format — Evidence tactic fields are now normalised to SCREAMING_SNAKE_CASE (e.g., "EMOTIONAL_MANIPULATION" instead of "Emotional Manipulation").

[1.0.0] - 2024-02-05

Added

  • Initial release of the Tuteliq TypeScript SDK
  • Safety Detection
    • detectBullying() - Detect bullying and harassment in text
    • detectGrooming() - Analyze conversations for grooming patterns
    • detectUnsafe() - Identify unsafe content (self-harm, violence, hate speech)
    • analyze() - Quick combined analysis with risk assessment
  • Emotional Analysis
    • analyzeEmotions() - Summarize emotional signals in content
  • Guidance & Reports
    • getActionPlan() - Generate age-appropriate action guidance
    • generateReport() - Create incident reports for professional review
  • Policy Management
    • getPolicy() - Retrieve current safety policy configuration
    • setPolicy() - Update safety thresholds and rules
  • Features
    • Full TypeScript support with comprehensive type definitions
    • Automatic retry with exponential backoff and jitter
    • Usage tracking via response headers
    • Typed error classes for different failure scenarios
    • Input validation for content length and message counts
    • Zero runtime dependencies (uses native fetch)

Security

  • API endpoint locked to official Tuteliq server
  • API key validation (minimum length, type checking)
  • Configuration bounds validation (timeout, retries)
  • No sensitive data exposed in error messages

[1.1.0] - 2026-02-10

Added

  • customer_id tracking field for multi-tenant / B2B2C scenarios
    • Available on all detection methods (detectBullying, detectGrooming, detectUnsafe, analyze, analyzeEmotions, getActionPlan, generateReport)
    • Echoed back in API response for easy correlation
    • Included in webhook payloads for routing alerts to the correct customer
    • Maximum 255 characters

[Unreleased]

Added

  • deleteAccountData() — Delete all account data (GDPR Article 17 — Right to Erasure)
  • exportAccountData() — Export all account data as JSON (GDPR Article 20 — Right to Data Portability)
  • AccountDeletionResult and AccountExportResult types

Changed

  • PII redaction is now enabled by default on the API (opt-out instead of opt-in)