The platform

One recognition layer.
Three connected systems. Always on.

Every frame you publish, every image your audience encounters, every scan that follows feeds a single system that gets sharper, faster, and more attributable over time. Three layers, one supply chain. From the surface your viewers see, to the recognition engine underneath, to the data you can export to your CRM.

Creating an IRCODE

From upload to live in minutes.

Drop in any image or frame and our advanced segmentation maps every object, region, and moment automatically. Tag what’s interactive, set the destination, and publish — no developer, no manual masking. The hard computer-vision work happens for you.

The architecture
01 · Lens SDK

The surface viewers see

White-label SDK inside the app you already own. Your brand, your scan brackets, your viewer experience. Flexible integration for iOS, Android, and Web.

02 · Recognition

The engine underneath

Patented image fingerprinting that matches the original frame in sub-second time. No QR codes, no embedded markers, no creative change required.

03 · Attribution

The data that lands

Every scan event is a first-party record, exportable to your CRM or warehouse. Campaign, daypart, geography, device, post-scan action. Deterministic. Exportable. Yours.

Layer 01Lens

The Lens SDK.

Your brand on top. Your viewers' phones in hand.

Lens is the surface layer of the platform. The white-label SDK that drops inside the app your audience already has on their home screen. Your brand colors, your scan brackets, your post-scan destination. The viewer never sees IRCODE. They open your app, raise their phone, and the experience you mapped is on screen in under a second. Flexible integration for iOS, Android, and Web. Lightweight SDK integration for your stack.

White-label SDK

Drops into iOS, Android, and Web. Your brand at the surface. Colors, scan brackets, animation, post-scan destination. Zero IRCODE chrome.

Sub-second recognition

Camera-to-match in under one second at couch distance. Works on live broadcast, streaming spots, OOH, packaging, and print.

Lightweight SDK

Flexible integration for iOS, Android, and Web. Your engineering team integrates one library. We run the recognition layer underneath.

Cross-surface coverage

One SDK recognizes commercials, billboards, magazine pages, packaging, stadium boards, and registered campaign assets. No per-surface tagging.

Live · KSL Salt Lake City · 71% of scans in-app
Branded scan experience
  • Custom scan brackets, color, and animation
  • Configurable post-scan destination (your app, your URL, your wallet pass)
  • Voice and haptic feedback hooks
  • Accessibility-tested camera and motion guidance
SDK integration
  • Native iOS (Swift) and Android (Kotlin) libraries
  • Web fallback for non-app surfaces
  • Sample apps and integration spec for engineering review
  • Flexible SDK integration for your stack
Asset registration
  • Self-serve registration portal for partner teams
  • Bulk upload for campaign launches and ad-flight rotations
  • Versioning and activation windows per asset
  • API for integration with existing DAM
Post-scan routing
  • Route to any in-app surface, web URL, or wallet pass
  • Conditional destinations by geography, time, and program
  • A/B test post-scan destinations across creative variants
  • Hand-off to commerce, loyalty, lead capture, or video player
Layer 02Recog

The Recognition Engine.

Image fingerprinting at broadcast scale.

The engine reads the frame itself. Not a QR code embedded on top of it. Not an audio watermark behind it. Not a category match against a generic visual taxonomy. Patented exact-match fingerprinting identifies the specific campaign asset on the screen. Even at couch distance, even on a phone held at an angle, even on broadcasts running in 52 countries. Sub-second response. Partner-scoped isolation between tenants. Deployable to match your stack.

Image fingerprinting

Patented exact-match recognition reads the original frame, not a marker layered on top. Zero creative modification required.

Real-time matching

Sub-second recognition from camera frame to attributed scan event. Optimized for live broadcast, streaming, and high-motion content.

Logical isolation

Every partner is a separate logical tenant, partner-scoped. We don't share scan signals across networks. We don't train models on partner data.

Global recognition mesh

Recognition infrastructure available in 52 countries with regional caching for sub-second response at the network edge.

Active · 1.5M+ scans recognized · 52 countries
Fingerprinting pipeline
  • Per-asset visual fingerprint generated at registration
  • Match across rotation variants and creative iterations
  • Patented exact-match, not category match, not similarity score
  • Continuous quality monitoring on production traffic
Recognition surface coverage
  • Linear broadcast and FAST channels
  • Streaming spots and CTV creative
  • OOH, stadium boards, packaging, and print
  • Live in-game footage and venue capture
Security and tenancy
  • Logical isolation between partner tenants
  • SOC 2 audit in progress
  • Per-tenant data partitioning and access controls
  • First-party data routing only. No cross-tenant aggregation.
Recognition reliability
  • Match confidence reported with every scan event
  • Fallback routing for low-confidence matches
  • 98.6% mobile-origination success rate in flagship deployment
  • Continuous A/B against new model revisions before rollout
Layer 03Attrib

The Attribution & Data Platform.

Where scan events become revenue.

Every recognition fires a deterministic event. The event is a first-party record you can export to your CRM. Campaign, program, daypart, geography, device, post-scan action, session depth, return cadence. No probabilistic models. No black-box mixed-media. The viewer raised their phone. The scan is the receipt. Your ad sales team walks into the renewal conversation with a number, not a media plan.

Per-scan attribution

Each scan event captures asset, campaign, program, daypart, geography, device, and post-scan action. Exportable to your warehouse or CDP.

First-party data ownership

Data is exportable to your CRM, not a shared marketplace. We don't sell or model partner scan data across tenants.

Advertiser reporting

Branded dashboards for ad sales teams. Per-advertiser, per-spot, per-flight. The performance story your reps need for renewal.

Session and cohort analytics

Track scan-to-conversion paths, return-user habit formation, and program-level performance to inform inventory pricing.

Reporting live · 56% iOS CTR · 17% return rate (KSL · 38 days)
Event schema and API
  • Deterministic per-event records, append-only log
  • Standard schema for asset, campaign, program, geo, device
  • REST and webhook APIs for downstream pipelines
  • Native exporters to Snowflake, BigQuery, S3
Advertiser dashboards
  • Per-advertiser performance views
  • Spot-level CTR, conversion, and session depth
  • White-labeled for client-facing reporting
  • Renewal-ready exports for ad sales teams
Audience and cohort tools
  • Scan-to-conversion path tracking
  • Return-user cohorts and habit-formation reporting
  • Program-level audience segmentation
  • Look-alike modeling for inventory pricing
Deterministic measurement
  • Direct-response attribution. No probabilistic models.
  • Per-spot ROI for local and national advertisers
  • Performance benchmarks across program and daypart
  • Tie-in with existing measurement (Nielsen, Comscore) as supplement
The scan pipeline

Five stages. Under a second.

Camera to matched asset, faster than the viewer can lower the phone. Recognition runs on the frame itself, against your registered catalog only, and hands off to the experience you mapped.

01

Capture

Frames are used for the match, then discarded — not retained.

02

Extract

Our proprietary recognition model produces an embedding from the frame itself.

03

Match

Exact Match runs the embedding against the partner-registered catalog. Only your assets, not a generic visual index.

04

Resolve

Under-a-second confidence resolution and post-scan routing lookup.

05

Deliver

IRCODE is invisible. The partner app receives the matched asset and routes to the experience the partner mapped.

Capabilities

Built for broadcast scale.

Recognition

Frames not retained

Frames are used for the match, not retained. An on-device offline mode is available.

Recognition

Real-time live broadcast

Recognizes the frame as it airs.

Recognition

Open-object

Works on any registered surface: frame, billboard, page, package, stadium board.

Vision

Multi-scale capture

Three-pass attention map for stability at couch distance and angle.

Security

First-party data isolation

Every partner logically isolated, partner-scoped.

Deploy

Under 200 lines of SDK

Lightweight SDK integration for your stack.

Data & privacy

Compliant by design. Yours, end to end.

SOC 2 Type II
In progress · target Q3 2026
GDPR compliant
EU data protection standard
No data resale
We never sell or repurpose your scan data
Patented
Patented recognition method
Bring it live

See the platform in action. Start with one network.

No full rebuild required. Pick the surface that hurts most. Local ad sales without attribution, a streaming library going dark, a sports rights deal renewing. We’ll scope a four-week pilot in the same call.

Talk to the team

Thirty minutes with the IRCODE team.

We'll show you exactly how the Lens turns your visuals into measurable, attributable channels, and how it plugs into your existing stack.