Vision Platform / VisionPROBE
01.5 / Vision Platform Concept Protects the signal

VisionPROBE

Consumer-edge monitoring at the viewing device — closing the chain.

VisionPROBE closes the monitoring chain at the consumer edge — observing player, network and device metrics from the actual viewing surface and correlating them upstream with VisionEDGE and VisionCLOUD for true source-to-screen visibility.

Concept / pre-product. Everything described here is a design target and 24-month roadmap intent — not shipped capability or measured performance.

Who it's for

For teams who need ground truth at the screen.

OTT / Streaming Operator

What actually reaches the screen.

Ground-truth of the delivered experience, not just server-side QoS — the last mile that upstream monitoring can’t see.

QoE / QoS Engineer

Telemetry from the consumption point.

Player, network and device metrics from the actual viewing surface, where degradation is felt.

Director of Engineering

Source-to-screen visibility.

Consumer-edge telemetry correlated upstream with VisionEDGE and VisionCLOUD for the whole path.

What it is

The viewer-edge pillar of the ecosystem.

VisionPROBE is designed to watch content where it is actually consumed — the viewer’s display, the terminus that VisionEDGE (production) and VisionCLOUD (in-flight) cannot see. A single Rust codebase is designed to auto-tier from ultra-light telemetry on constrained devices up to a full edge-intelligence node, delivered as SDK, browser extension, gateway agent or smart-TV integration.

Telemetry is designed to flow to VisionCLOUD, where VelaIQ correlates it for source-to-screen visibility. It is privacy-preserving by construction — metrics only, no content inspection on device, with one-way ACR fingerprinting. This is a concept solution: the sections below describe design intent and a 24-month, evidence-gated roadmap.

Design targetPre-product — no deployments or benchmarks yet
Single Rust coreOne codebase across four delivery channels
Privacy by constructionMetrics only, no on-device content inspection
Source-to-screenCorrelated upstream via VelaIQ
Capabilities

Design targets, not shipped capability.

The intended capability set for the consumer-edge pillar — presented as design intent.

01

Viewer-edge telemetry

Player metrics — buffering, bitrate switches, startup latency, time-to-first-frame — from the actual viewing surface.

PlayerStartup latencyTTFF
02

Network metrics

Latency, throughput, packet loss and jitter measured from the consumption point, not the origin.

LatencyLossJitter
03

Device metrics

CPU, memory, thermal and render statistics from the playback device itself.

CPUThermalRender
04

Adaptive probe

A single Rust binary auto-detects host resources at startup and self-configures to a tier, with safe upgrade and downgrade.

Auto-tierRust core
05

Privacy by construction

Metrics only, on-device PII filters, one-way ACR fingerprinting — audio hashed locally, opt-in for identity-linked telemetry.

Metrics-onlyOne-way ACR
06

Source-to-screen correlation

Telemetry correlated upstream with VisionEDGE and VisionCLOUD via VelaIQ — the delivered experience tied back to the signal.

VelaIQCorrelation

The core unit of value is the Ledge Event — a verified record of who, where, what device, what content and when. Census-scale across every opted-in viewer, rather than panel sampling.

What it monitors

A four-tier adaptive architecture.

One Rust core designed to scale from a heartbeat on a legacy set-top box to an on-edge analytics node — target envelopes below.

4
Adaptive tiers, one codebase
4
Delivery models
<0.5%
Target host-CPU overhead
TLS 1.3
Telemetry transport to VisionCLOUD
Tier 1 · UltraLite
  • <1 core, ~64 MB RAM
  • Heartbeat + basic player events
  • Legacy STB / low-end TV
Tier 2 · Standard
  • ≥2 cores, ≥64 MB
  • Player + network metrics
  • Modern STB / streaming stick
Tier 3 · Pro
  • ≥256 MB, ≥200 MB storage
  • Full quality + buffering forensics
  • High-end streamer / gateway
Tier 4 · Edge-Intelligence
  • ≥4 cores, ≥1 GB
  • On-edge analytics
  • Aggregates Tier 1/2 probes
The reporting surface · design target

A consumer-edge reporting library.

The intended report set for source-to-screen visibility — design targets.

Overview

Viewer-Edge Quality

Delivered-experience dashboard across the opted-in population.

Detail

Per-Probe Telemetry

Player, network and device detail from a single viewing surface.

Trend

Player Performance

Buffering, startup-latency distribution and bitrate-switch trending.

Region

Network by Region

Last-mile network performance mapped across geographies.

Correlate

Source-to-Screen

Consumer-edge telemetry tied back to VisionEDGE and VisionCLOUD.

Mix

Device & Platform

Tier distribution and device/platform mix across the fleet of probes.

Deployment & integrations

Four delivery models — all design targets.

Deployment

Pure software, no hardware capital — intended delivery across every viewing surface.

SDK / agent — iOS, Android, tvOS, Fire TV, Roku; JS SDK for web.
Browser extension — Chrome, Firefox, Safari, Edge.
Gateway / router agent — OpenWrt, RDK-B, no deep packet inspection.
Smart-TV native — via OEM partnership.

Integrations

Intended data path into the VisionCLOUD plane.

Rust core — to iOS, Android, tvOS, WebAssembly, Linux.
Transport — HTTPS / TLS 1.3, Protobuf or MessagePack.
Ingest — Kafka → TimescaleDB → stream processing for Ledge-Event composition.
VelaIQ correlation — with VisionEDGE and VisionCLOUD; REST API + webhooks.
Where in the chain

The consumer-edge node that closes the ecosystem.

VisionPROBE is designed as the third pillar — the consumer edge that completes source-to-screen visibility. Each Vision solution stands on its own; together they map the path the signal travels.

At the station

VisionEDGE

Monitoring where the chain begins.

Aggregation

VisionCLOUD

Cloud-scale rollup across the fleet.

Consumer edge

VisionPROBE

Verification at the point of delivery.

Status · pre-product

A concept on an evidence-gated roadmap.

No deployments and no benchmarks — every figure is a design target.

Status concept

Pre-product. Presented strictly as in-development design intent, never as shipped or measured capability.

Roadmap roadmap

A 24-month, evidence-gated path: Phase 1 MVP → Phase 2 Scale-Up → Phase 3 Platform.

Bring us a feed

Help shape VisionPROBE.

VisionPROBE is in design. If consumer-edge, source-to-screen visibility is on your roadmap, bring us the surfaces you need to see — and help shape where it goes.

01

Use-case scoping

02

Architecture review

03

Design partnership

04

Roadmap alignment

Start the conversation →

info@vela.com · (727) 507-5300