NexVDO — Edge AI vision by YUAN High-Tech
Edge AI vision, from camera to decision.
One platform for capture, video pipelines, and on-device AI — 30 years of video engineering, running on NVIDIA Jetson and Intel x86, included with YUAN hardware.
NexVDO unifies capture, recording, streaming, and on-device analysis through one SDK surface — the shift to edge AI that runs where the cameras are.
Trusted inside the ecosystems that matter
NVIDIA NPNPreferred Partner 2025YUAN High-TechTWSE: 5474Allxonfleet managementJetsonThor → Orin NanoDeepStreaminferenceTensorRToptimized inferenceNeteonU.S. front doorNVIDIA NPNPreferred Partner 2025YUAN High-TechTWSE: 5474Allxonfleet managementJetsonThor → Orin NanoDeepStreaminferenceTensorRToptimized inferenceNeteonU.S. front door
From camera to intelligence, in four modules
One SDK surface across NVIDIA Jetson and Intel x86. Each module ships with the same API discipline — and the same evaluation path.
A/01
Module overview
Capture
Bring physical video sources into one SDK surface, across Jetson and x86 targets.
A/02
Module overview
Record
See what never drops — 3 a.m. included.
DVR-grade recording shaped by three decades of capture engineering.
A/03
Module overview
Stream
Move live video between devices and services on production streaming paths.
A/04
Module overview
Analysis
Run vision models against live video on-device, with a documented custom-weights path (QDEEP).
0 yrs
of video-capture engineering
0M+
video products shipped annually
~0%
global capture-card market share
0
NVIDIA NPN Preferred Partner
Corporate figures: YUAN High-Tech (TWSE: 5474), as published 2026.
Quickstart
Working code in the first screen
Three checks in five minutes: read about it, see working code, know the vendor is alive.
- Free with YUAN hardware — zero licensing cost in the adoption decision
- Not open source — a supported evaluation path instead, answered in 48 hours
- GitHub-first support: issues answered by engineers
- Evaluation hardware ships from U.S. stock
quickstart.cpp
// Illustrative sketch — the full API surface ships with evaluation access
auto device = nexvdo::open_first_device(); // YUAN capture hardware
auto session = device.record("clip.mp4"); // Record module
device.stream_preview(display_frame); // Stream module
auto meta = session.analyze(model_handle); // Analysis (QDEEP path)// Sample repositories — pending publication (GitHub org transfer in progress).
One vision stack — from model weights to running systems
01
NexVDO SDK
The vision plane that feeds AI on-device: Capture, Record, Stream, Analysis. Included free with YUAN hardware.
Explore02
VideoAI productization
A documented, fixed-scope sprint that takes your model from weights to product. Process stages and results — published after approval.
Explore03
Systems & hardware
SmartVDO Edge systems on Jetson, the Pandora developer kit, capture cards, and the SmartVMS cross-camera tracking stack for greenfield deployments.
ExploreYUAN × NVIDIA — NPN Preferred Partner · 2025
Built inside the NVIDIA ecosystem
YUAN is an NVIDIA NPN Preferred Partner (2025). The SmartVDO Edge system line is built on Jetson — Jetson Thor to Orin Nano — with the Pandora developer kit alongside it. NexVDO rides the NVIDIA stack rather than replacing any part of it, and device and fleet management is answered through the Allxon partnership.
The Edge platform, Jetson Thor to Orin Nano
YUAN’s SmartVDO Edge processing platform spans 34 to 275 TOPS and handles multiple 4K60 channels for capture, recording, streaming, and analysis — with NexVDO included.
EDG6N0-S T5X
Edge system · Jetson Thor
Full specs at yuan.com.tw
EDG6N0-S T5X NVR
Edge system · Jetson Thor · NVR
Full specs at yuan.com.tw
VPP6N0-S AGX
Edge system · Jetson AGX Orin · 5G / PSE variants
Full specs at yuan.com.tw
VPP6N0-S-W NX
Edge system · Jetson Orin NX · PSE variant
Full specs at yuan.com.tw
EDG6N0-C-W ONX 4TVI DIO
Edge system · Jetson Orin NX · TVI + DIO
Full specs at yuan.com.tw
PND6N0 MB ONX
Pandora developer kit · Jetson Orin NX
Full specs at yuan.com.tw
Solutions
Where NexVDO lands
Six vertical workflows where cameras meet edge AI. Each pairs on-device analysis with YUAN Edge hardware and an evaluation path.
Buying a product?
You are building your own device — NexVDO SDK plus YUAN hardware inside your product.
Start at the SDK01Commissioning a system?
You need an integrated stack delivered and supported — hardware, software, and integration through Neteon.
Request scoping02Developer hub
Support that scales with your build
Every support surface answers from the same place: engineers who work on the SDK. Pick the channel that fits the question.
GitHub Discussions & Issues
Ask in the open. Issues and discussions are answered by engineers and stay searchable for the next team.
Open GitHuborg transfer in progress
Instant AI support
An assistant trained on the NexVDO documentation set, for answers at typing speed.
“Which Edge system fits a four-channel ingest?”
launches with docs
Get instant answerslaunches with docs
Documentation
Hardware and SDK documentation with step-by-step guides and honest compatibility tables.
View documentationin preparation
Quick answers
NexVDO is included free with YUAN hardware — there is no license line in the adoption decision. It ships through a supported evaluation path rather than a public download.
Request an evaluation. An engineer sends a substantive answer within 48 hours, and evaluation hardware ships from U.S. stock.
One SDK surface across NVIDIA Jetson and Intel x86 targets. Per-platform compatibility tables are pending publication.
Neteon fronts first-line support in U.S. hours on committed SLAs, with escalation into YUAN's Taipei engineering team.
Use it with your stack
Named integrations reflect the stacks NexVDO deployments run beside. Detailed integration notes — pending publication.
DeepStreaminference pipelinesTensorRToptimized inferenceNVENChardware encodeGPUDirectzero-copy videoJetsonedge computePyTorchmodel trainingTensorFlowmodel trainingONNXmodel interchangeOpenCVvision processingGStreamermedia pipelinesFFmpegtranscode toolingROS 2robotics middlewareC++native APIPythonscripting APIDeepStreaminference pipelinesTensorRToptimized inferenceNVENChardware encodeGPUDirectzero-copy videoJetsonedge computePyTorchmodel trainingTensorFlowmodel trainingONNXmodel interchangeOpenCVvision processingGStreamermedia pipelinesFFmpegtranscode toolingROS 2robotics middlewareC++native APIPythonscripting API
Need a human?
Engineers answer evaluation requests within 48 hours. Business and hardware questions route to Neteon — YUAN’s U.S. front door.
Evaluate on real hardware
Request an evaluation and get a substantive answer — from an engineer — within 48 hours. Evaluation hardware ships from U.S. stock.
Request an evaluation
- Engineer reply in 48h
- Ships from U.S. stock
- Free with YUAN hardware