NovaSDR

NovaSDR is a WebSDR server that streams waterfall/spectrum and demodulated audio to a browser UI.

• Backend: Rust (SoapySDR + OpenCL clFFT in the recommended build; CPU-only builds are supported)
• Codecs: waterfall zstd, audio flac
• Transport: WebSockets for real-time streams + HTTP for static UI

Community

👉 Join the forum: https://forum.sdr-list.xyz/

Showcase

What this repository contains

• crates/novasdr-server/: server, WebSocket endpoints, DSP runner
• crates/novasdr-core/: configuration, DSP primitives, codecs, protocol types
• frontend/: React UI (shadcn/ui), built as static files and served by the server
• docs/: maintained documentation (start at docs/index.md)

Installation

One-line installer (Linux/macOS)

Installer URL:

  curl -fsSL https://novasdr.com/install.sh | sh

The installer supports:

• Building NovaSDR from source
• Building and installing SoapySDR from source (always; recommended for direct hardware access)
• Installing OpenCL headers/runtime (optional)
• Installing Vulkan/glslang/SPIRV-Tools (optional; needed for --features vkfft builds)
• Building a SoapySDR device module from source (RTL-SDR, HackRF, Airspy, etc. where supported)

See docs/INSTALL.md for details.

Linux packages (manual)

NovaSDR runs either with:

• SoapySDR (recommended; direct hardware access), or
• stdin (pipe raw samples from an external capture tool as stdin), or
• fifo (pipe raw samples from an external capture tool as a file).

The sections below are copy/paste one-liners for a full source build (Rust + frontend + SoapySDR + OpenCL + clFFT).

Debian/Ubuntu (apt)

sudo apt-get update && sudo apt-get install -y --no-install-recommends \
  ca-certificates curl git tar \
  build-essential cmake pkg-config \
  clang libclang-dev \
  swig python3 python3-dev python3-numpy \
  nodejs npm \
  ocl-icd-opencl-dev ocl-icd-libopencl1 \
  libclfft-dev \
  libusb-1.0-0-dev \
  libopus-dev

Fedora (dnf)

sudo dnf install -y \
  ca-certificates curl git tar \
  gcc gcc-c++ make cmake pkgconf-pkg-config \
  clang llvm-devel libclang-devel \
  swig python3 python3-devel python3-numpy \
  nodejs npm \
  ocl-icd ocl-icd-devel \
  libusb1-devel \
  opus-devel

Arch (pacman)

sudo pacman -Sy --noconfirm --needed \
  ca-certificates curl git tar \
  base-devel cmake pkgconf \
  clang llvm libclang \
  swig python python-numpy \
  nodejs npm \
  ocl-icd opencl-headers \
  libusb \
  opus

openSUSE (zypper)

sudo zypper --non-interactive refresh && sudo zypper --non-interactive install -y \
  ca-certificates curl git tar \
  gcc-c++ make cmake pkg-config \
  clang llvm llvm-devel libclang-devel \
  swig python3 python3-devel python3-numpy \
  nodejs npm \
  OpenCL-Headers ocl-icd-devel \
  libusb-1_0-devel \
  libopus-devel

macOS (Homebrew)

brew update && brew install \
  git cmake pkg-config \
  llvm \
  swig python \
  node \
  libusb \
  opus

Note

If you see a bindgen error like Unable to find libclang, set LIBCLANG_PATH (for example on macOS: export LIBCLANG_PATH="$(brew --prefix llvm)/lib"), then rebuild.

Build (backend + UI)

cargo build -p novasdr-server --release --features "soapysdr,clfft"
cd frontend && npm ci && npm run build && cd ..

If you want stdin-only mode, omit soapysdr. If you want a CPU-only build, omit clfft.

Configure

Recommended: run the interactive wizard:

./target/release/novasdr-server setup -c config/config.json -r config/receivers.json

To edit an existing configuration with the same wizard:

./target/release/novasdr-server configure -c config/config.json -r config/receivers.json

Or edit config/config.json and config/receivers.json manually:

• config/config.json: server/websdr/limits + active_receiver_id
• config/receivers.json: receiver + input settings (sample rate, FFT size, sample format, signal type)

Run

./target/release/novasdr-server -c config/config.json -r config/receivers.json

Open: http://localhost:9002

SoapySDR mode (feature-gated)

With --features soapysdr, NovaSDR can open SDR devices directly from config/receivers.json (receivers[].input.driver.kind = "soapysdr"), or you can run novasdr-server setup to scan devices and generate receivers. In the recommended build we enable both soapysdr and clfft.

SoapySDR itself is a system library. For maximum compatibility, build SoapySDR from source:

git clone https://github.com/pothosware/SoapySDR.git
cd SoapySDR
mkdir build && cd build
cmake ..
make -j"$(nproc)"
sudo make install
sudo ldconfig # Debian/Ubuntu
SoapySDRUtil --info

Device examples

SoapySDR (recommended)

The simplest workflow is:

1. Build with --features "soapysdr,clfft"
2. Run novasdr-server setup and select a device (first launch will prompt automatically when config files are missing)
3. Run novasdr-server

If you prefer to write config/receivers.json manually, an RTL-SDR example looks like:

{
  "kind": "soapysdr",
  "device": "driver=rtlsdr",
  "channel": 0,
  "format": "cs16",
  "agc": false,
  "gain": 35.0,
  "gains": { "LNA": 30.0 },
  "settings": { "biastee": "true" }
}

stdin (pipe)

Make sure input.driver.format and input.signal match what the capture tool outputs.

RTL-SDR (IQ, u8)

rtl_sdr -g 48 -f 100900000 -s 2048000 - | ./target/release/novasdr-server -c config/config.json -r config/receivers.json

HackRF (IQ, s8)

hackrf_transfer -r - -f 100900000 -s 8000000 | ./target/release/novasdr-server -c config/config.json -r config/receivers.json

Airspy HF+ (IQ, s16)

airspy_rx -r - -f 648000 -s 912000 | ./target/release/novasdr-server -c config/config.json -r config/receivers.json

RX888 MK2 (real, s16)

rx888_stream -s 6000000 | ./target/release/novasdr-server -c config/config.json -r config/receivers.json

Documentation

Start here: docs/index.md

Key docs:

• docs/BUILDING.md
• docs/CONFIG.md and docs/CONFIG_REFERENCE.md
• docs/PROTOCOL.md
• docs/ARCHITECTURE.md
• docs/DSP.md, docs/AUDIO.md, docs/WATERFALL.md
• docs/OPERATIONS.md and docs/TROUBLESHOOTING.md
• docs/LICENSING.md and docs/THIRD_PARTY.md

The GitHub wiki is published from docs/ (see tools/publish_wiki.sh). On Windows, use tools/publish_wiki.ps1.

License

NovaSDR is licensed under GPLv3.

• LICENSE
• NOTICE

Attribution

NovaSDR is a continuation of PhantomSDR-plus (maintained by magicint1337) and includes contributions from NovaSDR contributors.

Copyright (c) 2025-2026, magicint1337.
Copyright (c) 2025-2026, NovaSDR contributors.

See NOTICE and docs/LICENSING.md for details and upstream attribution.