Flowbit System V1.5 Release Note

Flowbit System V1.5 Release Note

Added real-time 3D AI detection and segmentation, higher-accuracy multi-camera sync, and updated developer tools.

Added real-time 3D AI detection and segmentation, higher-accuracy multi-camera sync, and updated developer tools.

We have just released Flowbit System V1.5, the second major update to the Flowbit platform this year. Its predecessors V1.3 and V1.4 addressed issues like multi-camera fusion, data streaming, and cross-platform deployment. Flowbit System V1.5 builds on that work. It advances the platform's capability from capturing data efficiently to understanding the physical world in real time.

Spatial computing is not about moving the physical world into a computer. It is about helping computers understand the physical world. 

What's new

Real-time 3D AI detection and segmentation

  • Real-time capture and rendering at 30 FPS

  • End-to-end AI latency of 200 ms or less

  • Dual output of detection results in both camera coordinates and world coordinates

Traditional pipelines process images frame by frame. They answer only one question: what is in this frame? Flowbit answers a different one: what are the objects in the scene, and where does each one sit in 3D space? The AI inference runs separately from capture and rendering. Therefore, inference latency never blocks the data stream. When results do not return in time, masks can be reused across frames, leaving the rendering frame rate unaffected.

Multi-camera time synchronization with higher accuracy

The prerequisite for achieving 3D understanding is that all cameras run on synced timestamps. Flowbit System V1.5 rebuilds the multi-camera clock sync pipeline. It supports long, continuous capture with three or more cameras, without accumulated sync drift.

Sync offset, dropped frames, and mismatch counts are also exposed to developers as diagnostic outputs. This provides a verifiable time reference for multi-view fusion, point cloud stitching, and downstream spatial computing.

Deployment and developer kit

The platform is compatible with multiple 3D sensing technologies, including RGB, RGB-D, stereo, and LiDAR.

Supported platforms

SDKs have been released concurrently for Windows 11, Ubuntu 24, and Ubuntu 22. See details here: https://www.determtech.com/developer/docs-1.5/introduction.

Unity plugin

  • Wraps the Server SDK

  • Supports in-editor capture of synchronized multi-camera RGB and depth point clouds

  • Supports start/stop control of the 3D segmentation detector

  • Supports direct storage of recordings to the server 

Multi-camera calibration tool

  • Simplified the multi-camera calibration process. Users can now use just one mobile phone to assist the calibration of several cameras altogether.

  • Detection and segmentation results can be switched between camera coordinates and world coordinates.

Debugging and visualization tools

Front-end rendering tools and sample projects are now provided. They let users visualize live point cloud data and verify mask overlays.

SDK downloads and documentation are available in the Developer.

What's next

Flowbit System V1.5 is one milestone on our real-time spatial AI roadmap. Upcoming releases will expand to cover more object categories, a broader platform ecosystem, and deployment scale.

From data to intelligence, in a blink.

For media and business inquiries, contact contact@determtech.com.

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