Horizon Cloud Services
HORIZON MEDIALAKE · CONTENT INTELLIGENCE
ContentEye

Broadcast scale
intelligence.

A cloud-native, AI-powered content intelligence platform for automated media ingestion, deep analysis, quality control and compliance — built for large-scale broadcasters.

98.4%

ANALYSIS ACCURACY

<200ms

INGEST LATENCY

60+fps

ANALYSIS RATE

Request a DemoExplore Capabilities

Core analysis capabilities

Deep, multimodal analysis

ContentEye analyses video, audio, image, and document assets so broadcasters can automate quality checks and editorial validation at scale.

Every frame, seen

Real-Time Detection

Object, face, and scene detection with semantic context — down to the individual frame.

  • Objects, faces, scenes, and on-screen text in one pass
  • Semantic context, not just labels — who, where, and what’s happening
  • Frame-accurate output ready for QC and compliance
frame 84,221 · live
anchor_01 · 99.2%
logo · brand_x
scene · studio
text · lower-third
● objects · faces · scenes · text

Archives that stay current

Dynamic Metadata

Evolving metadata tags that reflect new topics and cultural trends, keeping archives current without rework.

  • Tags refresh as topics, events, and culture move on
  • Old footage resurfaces for new trends automatically
  • No re-indexing projects, no manual re-tagging
asset_0412.mxf · tagsNo re-indexing
On ingest · 2019
cricketstadiumcrowdsports
Auto-refresh · 2023
cricketIPL finallast-over thrillerviral moment
Auto-refresh · today
cricketT20 World Cupreels-readytrending audio
Tags refresh as topics and culture move — archives stay current

Content meets business data

Correlated Insights

Correlates in-content events with behavioural and business data for deeper editorial and commercial insight.

  • In-content events aligned with engagement curves
  • See which moments actually drive viewing and revenue
  • Editorial and ad decisions backed by evidence
In-content events × engagement+41% on goal moments
▲ goal
▲ red card
▲ winner
Editorial & commercial decisions backed by what actually happens in the content

Ask in plain language

Timeline-Synced Search

Instantly locate exact moments — key phrases, images, or shots — across massive archives with natural language.

  • Natural language queries across the entire archive
  • Results land on exact timecodes, not whole files
  • Phrases, faces, objects, and shots all searchable
“minister waves to crowd at airport”
evening_news_0704.mxf→ 12:41
archive_visit_2019.mp4→ 03:17
summit_recap.mxf→ 41:05
Natural language in — exact timecodes out, across the whole archive

Autonomous, with oversight

Agentic AI

AI agents autonomously manage validation, enrichment, SEO, and localisation — with human oversight for sensitive calls.

  • Agents run validation, enrichment, SEO, and localisation
  • Sensitive decisions pause for human review
  • Full trace of what each agent did and why
Agent run · overnight batch4 agents active
Validate subtitles · 42 assets✓ auto
Enrich metadata · sports archive✓ auto
Generate SEO descriptions✓ auto
Localise promos → 6 languages✓ auto
Archive takedown request · rights conflict⏸ human review
Agents handle the volume — people keep the judgment calls

Explainable by design

Compliance Automation

Enforces location-specific rules with explainable AI logs, consent tracking, and prioritised review queues.

  • Territory-specific rules applied automatically
  • Every decision logged with its reasoning
  • Consent tracking and prioritised review queues built in
asset_0079 · territory checksExplainable AI
IN✓ clear
EU⚠ GDPR: blur faces
US✓ clear

reason: face_id #221 has no EU consent record

rule: GDPR Art.6 · consent_registry.eu

action: auto-blur frames 1,204–1,377 → review queue

audit: trace logged · report #CE-8841

Every automated decision comes with its reasoning attached

See it in action

Every frame, analysed live

ContentEye scans each frame as media flows in — detecting objects, faces, and scenes, then attaching compliance-ready metadata in real time.

👤 Face detected · anchor_01🎬 Scene: newsroom, indoor🎙️ Speech: 98.7% confidence✅ Compliance: clear
newscast_0711.mxf · frame 10,412
☁️ Ingesting🧠 Analysing✅ Compliance Clear

Content Quality Check

Automated QC for every asset.

The Content Quality Check layer delivers an automated, AI-powered solution for validating technical and editorial quality across all media formats.

  • /01Extracts bitrate, resolution, duration & codec metadata
  • /02Detects blurs, black screens, frozen frames & audio issues
  • /03Validates speech-to-text accuracy, sentiment & subtitles
  • /04Flags inappropriate visuals and restricted text in images & documents
  • /05AI-prioritised manual review queues for sensitive assets
  • /06Actionable insight dashboards & compliance reports
CONSOLE_ROOT:~/qc_engine_v4

$ contenteye qc --asset newscast_0711.mxf

[03:14:02] INITIALIZING INGEST_STREAM_01..

[03:14:03] DETECTED: 24FPS // PRORES_422_HQ // 1080p / 12 Mbps

[03:14:04] ✓ METADATA EXTRACTED · duration 42:16

[03:14:05] ✓ VIDEO QC · no black frames, no freeze

[03:14:05] RUNNING FACE_ID · anchor_01 matched (99.2%)

[03:14:06] ⚠ AUDIO QC · loudness -21 LUFS (target -23)

[03:14:07] ✓ SUBTITLES VALIDATED · 98.7% accuracy

[03:14:08] ✓ VISUAL COMPLIANCE · clear

[03:14:09] ✖ COMPLIANCE_BREACH · logo_02 unauthorised territory DE

[03:14:10] LOGGING EVENT · frame 84,221 · report #FF-2921

[03:14:10] → Report: dashboards/qc/newscast_0711

_

Powered by Horizon MediaLake

A cloud-native AI backbone

Horizon's proprietary MediaLake is the microservices, serverless infrastructure that powers and scales ContentEye — from broadcasters to newsrooms to OTT streaming.

Scalable Architecture

Event-driven workflows, elastic scaling, and secure APIs.

Ingestion & Normalisation

Multi-source ingest, checksum integrity, automated format standardisation.

Metadata Enrichment

Codec analysis, segment tagging, and standardised metadata propagation.

Automated QC

AI-based anomaly detection across frames, audio, and text.

Compliance & Governance

Configurable rule-based filters with human review layers.

Search & Discovery

Natural language and semantic search with timeline syncing.

Workflow Orchestration

Real-time monitoring, notifications, and operational dashboards.

Data Security

Encryption, access management, and audit-enabled transparency.

Why ContentEye

Built for the signal, not the slide.

Serverless event-driven pipelines. Cloud-agnostic, enterprise-grade. Compliance-first, human-in-the-loop design. Standardised metadata across every asset in your archive.

Multi-format

Video, audio, image & document ingestion across every broadcast format.

Serverless

Event-driven pipelines that scale on demand with zero infrastructure management.

Frame-level

Granular insights and compliance indicators down to the individual frame.

Cloud-agnostic

Deploy across AWS, GCP, or Azure — no vendor lock-in.

Human-in-loop

AI-prioritised review queues for sensitive decisions requiring editorial judgment.

Explainable

Full AI audit logs with reasoning traces for every automated decision.

Territory-aware

Enforces GDPR, CCPA, and accessibility rules per region automatically.

Search-native

Natural language queries across massive archives with timeline sync.

Maximise ROI

Predictive optimisation of edits, distribution, and engagement.

Discoverability

Dynamic tagging and cross-linked archives reveal long-tail value.

Scale-ready

Elastic infrastructure with zero-downtime deployments at any volume.

API-first

RESTful APIs and webhooks integrate with any existing media workflow.

Licensing

Basic & Premier licensing models available

Empower your content to think — and your business to lead. Contact our sales team for licensing details and a personalised demo.

Contact Sales