Flowbit Server SDK

The SDK is a self-contained tree you copy to a workstation and build against. It captures synchronized colour and depth from the cameras, groups multi-camera frames by capture time, and optionally runs 3D object segmentation.

Language: C++ · Platforms: Windows x64, Ubuntu 22.04, Ubuntu 24.04

What the SDK is

The tree ships example sources, not prebuilt executables. The viewers and any application you write are compiled on the machine that run them against the exported package.

What's in the package

<prefix>/                       (this folder — call its path <prefix>

<prefix>/                       (this folder — call its path <prefix>

<prefix>/                       (this folder — call its path <prefix>

Choose the right platform tree

[!IMPORTANT] The compiled libraries are platform-specific, so the SDK is exported as three separate trees — Windows x64, Ubuntu 22.04 and Ubuntu 24.04 — and you must use the one matching your machine. The two Ubuntu trees are not interchangeable: their .so are built against that release's system libraries, most importantly GStreamer (22.04 ships 1.20, 24.04 ships 1.24) and glibc, so a 22.04 tree will not load on 24.04 or the reverse. The folder layout and instructions are identical for all three — only the compiled binaries differ.

Release and Debug builds

This export ships both Release and Debug builds of the SDK libraries. Build your application or the viewers in whichever configuration you want to link against — find_package(flowbit-server) picks the matching set per configuration automatically. The debug libraries carry a d suffix (for example flowbitserversystemd.dll) so the two sets never collide. A release-only export simply omits the …/debug/ folders.

Four libraries

Module

Provides

flowbitserver::flowbitserversystem

System, config and logging

flowbitserver::flowbitservercapture

CaptureClient, FrameSyncRunner, Raw_Data

flowbitserver::flowbitservermanage

Manager — device discovery, config, FTP

flowbitserver::flowbitserverdetect

Optional: 3D segmentation SDK

TO DO (software): One or two sentences per main type, and which header each lives in.

SDK prerequisites

NVIDIA CUDA runtime needs to match the build; GPU RTX 3060 or newer (compute capability ≥ 8.6)


Windows

Ubuntu 22.04 / 24.04

Compiler

Visual Studio 2019 or 2022

sudo apt-get install build-essential

CMake

≥ 3.18

≥ 3.18

GStreamer

MSVC x86_64, 1.18–1.24, with ...\1.0\msvc_x86_64\bin on the system PATH

libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev libgstreamer-plugins-bad1.0-dev libgstrtspserver-1.0-dev gstreamer1.0-plugins-{base,good,bad,ugly} gstreamer1.0-libav gstreamer1.0-tools

mDNS

Bonjour

libavahi-client-dev + avahi-daemon running

GL / X11

pkg-config libgl1-mesa-dev libglu1-mesa-dev libx11-dev

Python

3.10+ (detection only)

python3 python3-venv python3-pip (detection only)

[!IMPORTANT] GStreamer must be on PATH at runtime, not only at build time — the viewers load its plugins dynamically. On Windows, if ...\1.0\msvc_x86_64\bin is not on PATH, the viewer cannot decode the stream.

In this section

On this page

Let’s Build the Future of Physical AI Together

Let’s Build the Future of Physical AI Together

Let’s Build the Future of Physical AI Together

© 2026

Determ Technology · All rights reserved

contact@determtech.com