CameraStack

CameraStack
Server.

The software half of a network camera. Point it at your hardware and it handles the whole chain: sensor capture, hardware encode, RTSP streaming, ONVIF services, TLS security and device management.

In active development Evaluation builds available

How it works

One core.
Focused plugins.

Stop rebuilding the same software stack for every chipset. The differences live in focused vendor plugins; the product you maintain stays one codebase.

CameraStack Server architecture: ONVIF, WS-Discovery and RTSP interfaces into a portable core, which drives vendor adapters through a shared GStreamer porting layer or custom SDK integration, down to each vendor's sensor and SoC.

Start a product

You are building a new camera, encoder or NVR. Start from a working device instead of an empty repo: the whole video path from sensor capture and hardware encode to live RTSP streaming, plus ONVIF and device management, ported to your silicon.

Complete product stack

Complete a product

Your product already works and has its own video pipeline; it just needs a standards-compliant ONVIF interface. The CameraStack protocol layer bolts onto what you already have.

Drop-in interface

Built to travel

Highly
portable.

The hardware-specific code is deliberately small. A portable core does the standards work, and everything below it is a thin adapter, so a new platform is an integration, not a rewrite.

Core

Portable Go core

ONVIF, WS-Discovery, RTSP streaming, TLS security and device management live in one portable Go core. The same standards logic runs on every platform, carrying no vendor assumptions.

GStreamer

Shared GStreamer core

On any platform with GStreamer, capture and hardware encode run through a shared GStreamer core. A new board is described rather than rebuilt, so bring-up is a thin adapter over elements you already have.

SDK

Plug in without GStreamer

Where there is no GStreamer, a well-defined producer interface lets a native vendor SDK feed encoded video into the core. The producer is language-agnostic and typically written in C, straight against the vendor SDK.

The core

What's in
the core.

Everything an IP camera needs in one stack: streaming, video pipeline, ONVIF, security, analytics and camera control.

ONVIF services
Device, DeviceIO, Media, Media2 (Profile S and Profile T), PTZ, Events, Imaging, Analytics and Advanced Security
Discovery
WS-Discovery, so ONVIF clients find the camera on the network
Edge AI analytics
Object detection on the camera's own AI accelerator, published as tracked ONVIF scene metadata, with live model selection
RTSP streaming
Full range of ONVIF transports: TCP, UDP, HTTP tunnelling, WebSocket and multicast
Video
Multi-profile H.264 and H.265, with JPEG snapshots and rotation
TLS
HTTPS management and snapshots, RTSP over HTTPS, secure WebSocket, rtsps:// streaming and SRTP over UDP and TCP, with camera-held certificates managed through the ONVIF TLS Configuration Add-on
Media Signing
H.264 and H.265 video signed inside the encoded stream with a unique identity for each camera, enabled through ONVIF Media2
PTZ
Pan and tilt moves, presets and home position
Audio
A second RTP track, configured through ONVIF Media1 and Media2
Motion detection
ONVIF MotionRegionDetector polygon rules on a compact activity grid
Imaging
Brightness, contrast, saturation, exposure and more, wired to the ISP
OSD
Text overlays
Live encoder control
Frame rate, bitrate and GOP changes mid-stream, on-demand keyframes
Demand-driven encoding
Each encoder runs and consumes resources only while clients consume its profile
Metadata and events
Metadata streaming and events, including the ONVIF ImageTooDark tampering event

Evaluate it

Put it on
your bench.

CameraStack Server is in active development ahead of its commercial launch. Evaluation builds are available to teams building real products, and if your chipset is not ported yet, the port can usually be folded into an integration project.

Request an evaluation build