Vision Platform / VisionCAM
01.3 / Vision Platform In development Protects the signal

VisionCAM

Multi-camera fleet health and frictionless contribution.

A single pane for the health of every camera and contribution feed — real-time multiviewer, per-camera quality and alerts, a fleet-wide alert timeline, and a sub-minute contributor on-ramp. Ingests dozens of IP cameras per commodity workstation; self-hosted and air-gap capable.

Pre-revenue, in active development with a design-partner cohort. Capabilities below separate what ships today from what is in development.

Who it's for

For multi-camera and live-event operations.

Live-Event / Multi-Camera Operator

Every camera on one view.

Each camera and contribution feed on a single wall, worst-first, with click-through from any alert straight to the camera.

Broadcast Engineer

Per-camera drill-down.

Continuous bitrate, frame-rate and quality sampling on every feed, with a severity-bucketed fleet alert timeline.

Production Lead

Visitors live in under a minute.

Generate a single-use invite; a contributor scans a QR code and is on-air via browser or RTMP in seconds.

What it is

One pane for every camera and contributor.

VisionCAM ingests, controls, monitors and queries fleets of IP cameras on one commodity x86 workstation — five ingest on-ramps (RTSP, SRT, HLS, WebRTC, RTMP) converging on a single operator workflow. A per-camera supervisor with an explicit restart policy keeps feeds up, and the dashboard rolls health across the last 24 hours, worst-first, with supervised recovery: the platform detects, the operator confirms, the platform recovers.

Frictionless contribution is the co-lead: an operator generates a slot, a contributor scans a QR code or follows a link, and they are on-air via WebRTC or RTMP in seconds — through single-use, time-boxed, audited invite codes. Underneath, a Schema Intelligence Engine learns each deployment’s structure from observation, so VelaIQ’s natural-language queries adapt without hand-authored schemas. The platform is self-hosted and air-gap capable, with no internet required.

Self-hosted, air-gap capableOne Linux box, no cloud, no internet required
No GPU requiredFor monitoring and contribution
Five ingest on-rampsRTSP · SRT · HLS · WebRTC · RTMP
Sub-minute QR-to-liveSingle-use, time-boxed, audited invites
Capabilities

Fleet health and frictionless contribution.

The real product: every camera watched, every contributor on-air in seconds, every action auditable.

01

Unified multiviewer

Every camera and contribution feed on one wall, with per-camera video statistics and alarm state.

60+ camerasOne wall
02

Fleet-health monitoring

Per-camera supervisor with restart policy; 24-hour roll-up, worst-first; continuous bitrate, frame-rate and quality sampling.

Worst-firstSupervised recovery
03

Fleet-wide alert timeline

Severity-bucketed alerts — critical, warning, info — with click-through from any alert to its camera.

Severity bucketsClick-through
04

Frictionless contribution

Single-use invite → QR or link → on-air via WebRTC or RTMP in seconds; browser, Android app or professional encoder.

BrowserAndroidRTMP
05

VelaIQ

Plain-language questions return a tabular result with a forensic proof receipt — a deterministic planner that refuses rather than fabricates.

Natural-languageProof receipt
06

Forensic audit trail

Every operator action and query captured in the same transaction as the write — a system of record without a separate compliance product.

Append-on-writeSIEM-readable

Capabilities are honest about stage: control plane, five-on-ramp ingest, contribution, VelaIQ and the audit trail ship today; SDK bindings, a live-state WebSocket hub, recording and frame-level quality monitoring are in development.

What it monitors

Built for density on commodity hardware.

Dozens of cameras per box on a single self-hosted workstation — no GPU required for monitoring and contribution.

5
Ingest on-ramps into one session
3
Contribution paths — browser, Android, RTMP
<1 min
QR-to-live for a new contributor
Air-gap
Capable — no internet required
Ingest

Five on-ramps, one session.

Every contribution and camera protocol converges on one operator workflow.

Camera ingest
RTSPSRTHLS
Low-latency
WebRTCRTMP
Contribution
BrowserAndroid appPro encoder
Substrate
PostgreSQL 16Object storage
The operator interface

One operator dashboard.

A single workflow surface — wall, health, alerts, contribution and query.

Wall

Camera Grid

Every camera and contribution feed, filterable by source and state.

Overview

Health Dashboard

24-hour roll-up, worst-first, with per-camera drill-down.

Triage

Alert Timeline

Severity-bucketed events with click-through to the camera.

On-ramp

Contribution Surface

Generate single-use invites and watch contributors come on-air.

Query

VelaIQ Console

Natural-language questions with a forensic proof receipt per answer.

Record

Audit Log

Every action and query, append-on-write, readable by external SIEM.

Deployment & integrations

One modest box, air-gap capable.

Deployment

A small container set on a single commodity x86 workstation — no cloud required.

One x86 server — multi-core CPU, 32 GB RAM, NVMe, Ubuntu 24.04, single monitor.
Four container images — live stack ~344 MB; cold-start under 15s.
Air-gap capable — image tarballs, no registry calls, customer-runnable acceptance script.
No GPU required — for monitoring and contribution.

Integrations

Built to slot into existing recording, MAM and control systems.

REST API — control and config (shipped).
WebSocket — live state (in development).
JWT + scoped tokens — argon2id, per-scope API access.
OpenAPI 3.1 — Go SDK shipped; TypeScript, Python, C#, Java in development.
Where in the chain

The contribution node of the Vision ecosystem.

VisionCAM is where cameras and contributors enter the ecosystem — at the venue and in the field. Each Vision solution stands on its own; together they map the path the signal travels.

Field & venue

VisionCAM

Multi-camera fleet health and contribution.

At the station

VisionEDGE

Monitoring where the chain begins.

Aggregation

VisionCLOUD

Cloud-scale rollup across the fleet.

Status · honest about stage

In active development.

Pre-revenue, with a design-partner cohort. Presented as what it is.

Shipping today

Control plane, five-on-ramp ingest, contribution, VelaIQ, audit trail, Go SDK and the air-gap showcase.

In development roadmap

SDK bindings, WebSocket hub, React UI, recording and frame-level quality monitoring.

Bring us a feed

Put VisionCAM on your camera fleet.

Engagements start with a signal. Bring a representative set of cameras and contribution feeds and we run them through the audit pipeline — you see exactly what VisionCAM does, before any commercial conversation.

01

Feed assessment

02

Platform audit

03

Operator review

04

Deployment scoping

Start with a signal →

info@vela.com · (727) 507-5300