NexVDO SDK

Build video AI from capture to inference.

One production toolkit unifies video capture, recording, streaming, and AI analysis across NVIDIA Jetson and Intel x86 systems.

Discuss Your ProjectExplore Capabilities
Unified API

C/C++ · Python · Qt

  1. 01

    Capture

    Video, audio, VANC

  2. 02

    Record

    Encode, mix, preserve

  3. 03

    Stream

    Move, control, deliver

  4. 04

    Analysis

    Detect, search, reason

Hardware-accelerated media and AI runtime

Developer interface
C/C++, Python, and Qt
Compute targets
NVIDIA Jetson and Intel x86
Architecture
One API across four modular workflows

Capability architecture

One SDK. Four connected workflows.

Compare video inputs, media engines, protocols, and analysis services at a glance, then follow the shared acceleration stack to the target platform.

API surfaceYUAN cross-platform unified API

Capture

Video ingest

  • 2D/3D video
  • Audio & VANC
  • USB cameras
  • IP streams
  • Virtual desktop

Rendering & imaging

  • High-performance renderer
  • Image snapshot
  • Mirror & clone
  • Region & HDR
  • Surround & multiview

Image processing

  • Deinterlace engine
  • Alpha blending
  • Scale & crop
  • Chroma key
  • RGB repack

Workflow control

  • Auto signal detection
  • Continuous snapshots
  • Format-change events
  • GPUDirect

Medical & image formats

  • DICOM
  • HL7
  • Worklist
  • BMP
  • JPG
  • PNG
  • TIF

Acceleration stack

  • GPUDirect / RDMA
  • NVENC
  • NVDEC
  • NPP
  • CUDA
  • TensorRT
  • TensorRT-LLM
NVIDIA Jetson + Intel x86

Integration model

A direct path from API to deployment.

The software stack keeps application logic consistent while the media and inference layers use the acceleration available on the target platform.

  1. 01

    Build with familiar tools

    Use C/C++, Python, or Qt through one unified API surface.

    Next stage

  2. 02

    Compose only what you need

    Connect capture, recording, streaming, and analysis as modular services.

    Next stage

  3. 03

    Use the available acceleration

    Map codecs, media operations, and inference to NVIDIA or Intel compute.

    Next stage

  4. 04

    Deploy on the target platform

    Move from prototype to YUAN edge systems without rebuilding the pipeline.

Application fit

Designed for video-intensive systems.

Use the same core pipeline when the project moves between live production, embedded inspection, clinical video, and remote operation.

Broadcast and live production
Synchronize capture, multi-track recording, compositing, and standards-based delivery in one chain.
Industrial inspection
Keep acquisition, low-latency transport, and visual inference close to the production line.
Smart healthcare
Connect clinical video workflows with DICOM, HL7, Worklist, and AI-assisted analysis.
Remote operations
Combine PTZ control, WebRTC or RTSP transport, recording, and analysis for distributed teams.

YUAN USA

Bring us your video pipeline.

Tell our U.S. team what you need to capture, move, and analyze. We will help map the software services and YUAN hardware to the project.

Contact YUAN USA