Changelog

Recent updates to Mediavox APIs and Developer Portal.

Latest 2026-07-16

CISO Threat Dashboard: Standalone dashboard at /brand-protection/dashboard. KPI cards (threats, active, takedowns, avg detection), timeline chart (30d), threats grid with actions, and export blocklist (hosts/Pi-hole/Cloudflare Gateway format).
Interactive Threat Graph: vis.js force-directed network showing domains, IPs, registrars, and IOCs as connected nodes. Reveals organized campaigns — not isolated domains. Fullscreen mode + PNG export. Available on Protect and Enterprise plans.
Content Forensics (inline in /urls/analyze): 7 deterministic signals analyzed on every URL: Telegram/Discord exfiltration, credential forms, brand font impersonation, suspicious JS, free hosting redirect, social engineering lures. Zero AI cost. Patterns configurable from BD.
Campaign Fingerprinting: IOCs (Telegram bot tokens, hosting paths, JS filenames, CSS fonts) automatically extracted from confirmed phishing. Future domains from the same attacker match instantly via fingerprint cache.
3 New CISO Endpoints: GET /bi/summary (KPIs), GET /bi/timeline (chart data), GET /bi/graph (nodes + edges). Multi-tenant IDOR protection.
Visual Verification Feature Flag: Headless browser screenshots gated by plan (Protect/Enterprise only). Configurable per-plan in BD — zero deploy to activate.
Domain Scanner: Now scans 3,000+ variants per domain (up from 300). Forensics results come from /urls/analyze response (centralized, all callers benefit).

2026-07-12

CT Log Real-Time Monitor: Persistent WebSocket connection to global Certificate Transparency stream. Detects typosquatting domains the instant they obtain an SSL certificate — before the attack launches. 349 LATAM brands monitored 24/7.
Abuse Report Generation: POST /threats/{id}/generate-abuse-template — Resolves hosting provider from IP, finds abuse contact email, generates formal DMCA/phishing complaint. Optional LLM enhancement for professional tone.
One-Click Takedown: POST /threats/{id}/send-abuse-report — Sends the abuse report to the hosting provider with one confirmation click. 24h dedup prevents duplicate reports. Tracks sent status in threat metadata.
Provider Abuse Database: 17 hosting/registrar abuse contacts indexed (Cloudflare, Hostinger, Namecheap, GoDaddy, Amazon, Google, OVH, Hetzner, and more). Auto-resolves from domain IP.
MFD Hardening: Fixed subdomain bypass in IsSuspiciousDomain (GetApexDomain), DNS thread starvation (CancellationToken 3s timeout), Regex ReDoS protection (compiled + cached + 1s backtracking timeout).

2026-07-11

Per-Empresa AI Configuration: Each empresa can now configure AI tone, response length, language, restrictions, and context. Applied to all chat products (Vixo, SC, HP). Zero deploy — 1 config change in BD.
MFD Precision Improvement: Fraud detection precision improved from 43% to 100% on real SMS samples. Brand shortener whitelist, typo detection, anti-poisoning, replay cache.
Shared Library (Mediavox.Common): All AI calls centralized in one shared library with triple fallback, telemetry, and model rotation from BD. Zero duplicated implementations.
Embeddings API: GenerateEmbeddingAsync — Gemini → OpenAI fallback (768 dimensions). Used by RAG indexer.
PgHelper Async: CallFunctionAsync, CallOperationAsync, CallScalarAsync available in shared library for all projects.
Dependencies: Google.Apis.Auth 1.73, Scalar 2.16.11, Newtonsoft.Json eliminated from taskmanager, dead MailKit removed.

2026-07-10

La Bóveda (DocumentPower): Persistent document vault. Every analyzed doc is stored with full-text search (Spanish), automatic deduplication, and storage quota per plan. Endpoints: /vault/browse, /vault/search, /vault/{id}, DELETE /vault/{id}
AI Provider Monitoring: Automatic health checks every 12 hours with instant email alerts if any provider goes down. Zero manual monitoring needed.
Centralized AI Configuration: Rotate AI models across all services with a single configuration change. No downtime, no redeployment.
Enhanced AI Resilience: Verified triple-provider coverage on all endpoints. Automatic failover ensures zero downtime if any single provider goes offline.
Degradation Detection: Proactive alerting when AI quality degrades — before it impacts users.
Infrastructure Monitoring: All 6 services now monitored with automatic alerting (previously only 2).

2026-06-28

HealthPower: 28+ endpoints (appointments, specialists, patients, consent, adherence, authorization, RIPS, risk scoring, anomalies, reconciliation)
Adherence monitoring: treatment registration, dose tracking, cascading alerts (patient → caregiver → auditor)
Validation Gateway: local authorization rules (auto_approve, copay, manual_review, quota enforcement)
AI resilience: automatic triple-provider failover (Gemini, Claude, OpenAI) for all text and document generation
Multimodal analysis: image and PDF processing with automatic provider failover
n8n node v1.7.1: 20 HP operations, 10 workflow examples, 6 products
DocumentPower: 18 document types, tipo hint parameter, file_url support with SSRF protection
Sales Copilot: GET /api/v1/sc/vendors endpoint for integrations

2026-05-18

All Turing AI endpoints standardized to /api/v1/*, consistent with Vixo API and mediaAPI
Official n8n community node: 8 resources, 57 operations (DataTools, Security, Recognition, Finance, AI, Turing, DocumentPower, Vixo API)
Pricing page cached (1h TTL): faster load, fewer DB queries
Widget v2.5 updated to use versioned API routes

2026-05-16

Turing AI: 120+ tools across all Vixo modules
Document follow-up: upload a file and ask questions about it in the same session
Multi-provider resilience: Claude fallback with full conversation history
Sector intelligence: 8 LATAM industries with auto-updated market data (TAT Retail, BPO, Distribution, Professional Services, Construction, Manufacturing, Logistics, Insurance)
Consultative selling: Turing uses sector data to personalize responses for prospects
Widget v2.5 with improved session persistence
DocumentPower: cross-reference uploaded documents with system knowledge

2026-04-25

Turing AI function calling: 120+ dynamic tools across all Vixo modules, AI Flywheel auto-learning
Turing Personal B2C: signup, login, dashboard, chat, policy upload with AI Vision
Treasury module: bank accounts, movements, reconciliation with state machine
51 LATAM banks (CO/MX/PE/CL) with SWIFT codes
Developer ↔ Vixo account linking (auto + manual)
Mobile QR session for Vixo AI chat
API Keys management page with product filter

2026-04-20

2026-04-19