Apply for Beta

For teams building radar sensing products

Build with radar like you have an expert on the team.

You have a radar dev kit on the bench and a sensing feature on the roadmap. RadarFlow helps your team connect supported hardware, understand live behavior, build the processing chain, and capture traceable evidence—without weeks of SDK plumbing or one-off infrastructure.

Apply for Beta

Private beta for teams building real radar products.

Spatial + Range-Doppler Raw echoes to a live range-Doppler map and spatial point cloud — running on a Pi 5, in your browser.
Micro-Doppler A fan's blade signature resolved live in the micro-Doppler spectrogram.
Vital Signs Phase-unwrapped displacement tracing breathing motion in real time.
Beamforming Azimuth and elevation cuts pinpointing targets in space.

Why It Matters

A demo is not a product decision.

You know radar is the right sensor: presence detection, vital signs, gesture recognition, no camera. But turning that into a shipped product normally takes a team of radar specialists. Your team is still weeks from a first trustworthy range-Doppler map, and the deadline keeps approaching.

Radar looks great in the vendor demo videos. But that's their scene and their config, not your enclosure, your mounting, your clutter. And it leaves you processed output, not the reproducible, labeled captures that become your training data when the product needs edge AI.

How It Works

From dev kit to live radar data in three steps.

Connect

Step 1
  • Flash the RadarFlow image to a Raspberry Pi 5, plug in your supported radar board, and open your browser.
  • No SDKs, no toolchains, no driver archaeology.

Build

Step 2
  • Drag processing blocks onto the canvas — FFTs, filtering, CFAR, detections — and wire them together.
  • Tune parameters live and see the effect immediately.

Capture

Step 3
  • Watch live range-Doppler displays as you build.
  • Record traceable HDF5 evidence your team can inspect and replay.
Two-Minute Build The entire build, unedited — empty canvas to a live range-Doppler display, shown at 6× speed. Click to watch the full build (1 min).

What Your Team Gets

From live sensor to decision-ready, AI-ready data.

Hardware Evaluation

Sensor
  • Bring up supported hardware and see live range-Doppler maps, traces, and spectrograms without rebuilding vendor workflows.
  • Tune configurations with the capture context attached, then make a defensible go/no-go decision.

Visual Radar Signal Processing

Pipeline
  • Build from raw sensor data through FFTs, filtering, CFAR, detections, and tracking.
  • Tune live or replay captures through new configurations without rerunning every bench test.

Validation Data Vault

Data
  • Record synchronized HDF5 with quality metrics, labels, and full configuration context.
  • Search, replay, share, and export the same data for validation, algorithms, and edge ML.

AI-Ready Data

AI
  • Turn validated captures into training-ready datasets with labels and full configuration context.
  • Build toward inference models with radar data your team can inspect, replay, and build on.

Built by radar engineers · 30 years across hardware, DSP, and systems · Why we built RadarFlow

Private Beta

Ready to put that radar expert to work?

Join the private beta and evaluate radar with real data—evidence your team can inspect, reproduce, and build on. We onboard a small cohort of teams at a time, so every team works directly with the engineers building RadarFlow.

Apply for Beta

Applications reviewed weekly.

What beta teams get

  • Every supported board — AKM today; TI, Infineon & Novelda in development
  • Advanced radar signal-processing blocks and application pipelines
  • Dataset Vault with searchable catalog, tags, and quality metrics
  • Full HDF5 export — your data is always yours

You'll need a supported radar board and a Raspberry Pi 5 — see supported boards.