Tagged: AI

SDR# VDK: Create Your Own SDR# Plugins by Prompting an AI Coding Agent

Over on the Airspy groups.io forum, Youssef (prog), the developer of SDR#, has announced the release of the SDR# VDK (Vibe Development Kit). The VDK is meant to let anyone create SDR# plugins with AI coding agents like Claude Code or ChatGPT Codex, without needing to be a programmer.

The kit includes a working plugin template, the SDR# API documentation, DSP primitives, and a set of instructions that teach the AI agent the coding practices Youssef uses in his own plugins. You unzip the VDK on a Windows PC, point the agent at the folder, and describe the plugin you want. The agent then writes the code, builds it and deploys the plugin into your local SDR# install.

To demonstrate it, Youssef had an AI agent produce a working DRM receiver plugin from a single prompt. In a second video, ChatGPT built a DAB/DAB+ plugin in about 20 minutes. Youssef notes that AI-written code still needs testing. The VDK is available as a zip file from the Airspy website, and plugins built with it require the latest SDR# build.

Forum members have already started sharing what they've made. Loyd Van Horn (W4LVH) from DX Central has released a clock plugin on GitHub that shows UTC and/or local time, and adjusts its layout depending on where it is docked. A user called bear has made an auxiliary VFO plugin called 'SecondVFO', and multiple copies of it can be run at once. PW00X has also shared a modified version of the DAB/DAB+ demo plugin, which adds a favorites list and automatic tuning.

Other users have posted results but haven't released their plugins yet. Edward MacDonald built an MCP server and connector plugin that lets Claude control SDR# directly, view the spectrum and work with the I/Q and audio streams. As a test, Edward asked Claude to find all local FM broadcast stations and list their names, frequencies and SNR values. Craig_H created an L-band ACARS decoder for the 600, 1200 and 10500 bps channels with a single Claude prompt, and TheP111 reports getting a working RTTY decoder plugin out of Claude Code in 15 minutes. Roberto Cipriani has replaced the old 32-bit SimpleDMR plugin with a VDK-built version and plans to do TETRA next. Youssef himself is currently using the VDK to write a DVB-T/T2 decoder from scratch.

We gave the VDK a test ourselves and had it code up a wideband FFT sweeper by giving Claude Opus 5.5 a simple plain english prompt "Make a plugin that allows us to view a wider spectrum, by rapidly sweeping across the spectrum". Claude prompted us to install the .NET 10 SDK, and then it created, compiled, and installed the plugin all on its own, giving a working result after only 10 minutes.

SDR# VDK: Wideband FFT Sweep implemented by Claude Opus 5.5 in 10 minutes.
SDR# VDK: Wideband FFT Sweep implemented by Claude Opus 5.5 in 10 minutes.
Airspy VDK Demo - DRM Plugin

SDR# VDK One Liner Demo

AI Disclaimer: The VDK is designed for building SDR# plugins with AI coding agents, and the plugins mentioned in this post were created with AI agents.

NightDX: Automated Overnight Shortwave DX Listening with an Airspy HF+

Thank you to Marco Zennaro for writing in and sharing NightDX, a free terminal app for macOS, Windows and Linux that lets an Airspy HF+ do your overnight shortwave listening for you. The menu interface is deliberately styled like early-90s DOS software.

NightDX pulls the EiBi, AOKI and HFCC broadcast schedules, plus live solar flux and Kp from NOAA. A propagation model then predicts what should reach your location and calibrates itself against what your own radio hears. The HF+ sweeps all the bands in about 10 seconds to see which carriers are really there, and overnight NightDX records the best stations, putting ones you've never logged first.

Optional AI features can then label each recording as speech, music, or noise using an AudioSet classifier, and transcribe and translate it with OpenAI's Whisper, which helps when writing reception reports. In the morning, a browser report shows the recordings with audio players, one-click reception reports drafted in Gmail, and a QSL card log.

We note that NightDX doesn't claim to identify stations. It shows schedule names as candidates and flags recordings where the evidence doesn't fit, such as the wrong language or a signal far too strong for the predicted path. Only the Airspy HF+ (Discovery or Dual Port) is supported for now, so there is no RTL-SDR support. Marco notes that other SDRs such as the Perseus and Elad are being looked into. The code is open-source under the MIT license and available on GitHub.

NightDX Screenshots
NightDX Screenshots

AI Disclaimer: Marco notes that NightDX was built with the help of Claude.

Claude VOACAP Skill: Ask Claude About Live Ionospheric Propagation Predictions to Optimize DX

Thank you to Reid Crowe for submitting his project, "voacap-skill." This skill for Claude Code AI (a frontier LLM) lets Claude access and use the latest VOACAP ionospheric propagation data in conversations.

Instead of manually filling out a form for each prediction and analyzing the results yourself, with the voacap-skill, users can ask Claude questions like, "What's the best band to reach Tokyo from Columbus, Ohio around 0300Z in October?" and Claude will run a VOACAP prediction in the background and return an answer. Reid explains:

If you've ever used VOACAP, VOACAP Online, or watched a contest DXpedition plan its band schedule, you already know the engine underneath this: it's the same NTIA/ITS point-to-point HF prediction model that's been the industry standard for propagation forecasting since the 1980s. It takes a transmitter location, a receiver location, a month, and a sunspot number, and tells you — for every hour of the day and every frequency you care about — how reliable that circuit is likely to be.

voacap-skill takes that engine (via voacapl, the modern Linux/GFortran port, on macOS/Linux, or the native NTIA/ITS Windows engine on Windows) and wires it into Claude Code as a "skill" — a capability Claude can reach for on its own.

The skill is free and open-source, with the code available on GitHub. Below, Reid provides some more examples of what you can ask Claude once it has access to VOACAP data.

AntSDR T510 Pre-launch: A 1 MHz to 6 GHz SDR with 2 GHz Bandwidth and a Built-In NVIDIA Jetson

MicroPhase has just announced the pre-launch of its AntSDR T510 AI software-defined radio platform. This high-end device combines the AMD Zynq UltraScale+ RFSoC ZU47DR with an NVIDIA Jetson Orin Nano. The AMD ZU47DR RFSoC alone is already a beast, covering 1 MHz to 6 GHz, with 14-bit ADCs and 8 RX and 8 TX channels. This is a direct-sampling SDR, meaning no tuners are used, and the ADC simply samples at an extreme 5 GSPS. After the DDCs, each channel has a maximum bandwidth of up to 2 GHz, but smaller channelization can easily be achieved within the large FPGA embedded in the RFSoC.

Obviously, processing a large amount of data requires powerful computing, so the AMD ZU47DR RFSoC is combined with an NVIDIA Jetson Orin NX for real-time AI signal processing and intelligent spectrum analysis.

MicroPhase also notes that the AntSDR T510 AI will be open source, allowing developers to easily work with the hardware. It will ship with an Ubuntu 22.04 system preconfigured with CUDA, include GNURadio and SoapySDR compatibility, and include GPU-accelerated demos for real-time Wi-Fi, Bluetooth, and modulation classification using RF data captured by the platform.

This is a high-end SDR comparable to Ettus products that cost US$30k+ like the USRP X410, but Chinese manufacturers like MicroPhase may be able to bring the pricing down. The ANTSDR-T510 non-AI version without the Jetson is already available on Chinese marketplace sites for about US$5000, so with the additional $500 Jetson Orin Nano added, this version should only be a little more expensive.

The AntSDR T510-AI
The AntSDR T510-AI

Scanner Command: Mapping Incidents from Live Radio Calls and ADS-B to Airband Voice with AI

Recently, we posted about DeepSDR, which is a program that automatically transcribes voice data from public safety radio communications and uses an LLM to categorize and plot incidents on a map as they happen.

In that post, we noted that we expect to see more projects like this soon, and sure enough, another program called "Scanner Command" with a similar angle has been submitted to the blog by author Benjamin Blood. Like DeepSDR the software listens to public safety and then uses voice transcription and an LLM to automatically understand, categorize and map incidents on the map. One additional feature that Scanner Command has is a fun tool that correlates ADS-B data with airband voice. Benjamin explains it best:

Scanner Command turns a radio scanner into a live AI incident map. A Uniden SDS200 covers the trunked digital public safety side (the LA area's P25 Phase II system, with ProScan driving the scanner and handing off every recording), and two $45 RTL-SDR dongles handle the rest, one decoding ADS-B with readsb, one capturing VHF airband transmission by transmission. Everything gets transcribed locally with Whisper, an AI extraction pass works out what happened and where, and about 30 seconds after a call ends it's a color-coded pin on the map with the audio one click away. Location matching is honest on purpose: an exact intersection match gets a solid pin, an area-only match gets a dotted one, and it tells you which.

Here's the design idea underneath it, and what I think separates it from the other transcribe-and-map projects out there: this is built for tactical situational awareness when the grid's down. The system does its learning while it's online, building its own gazetteer of local streets and places, learning unit callsigns and talkgroup names, priming the transcriber's vocabulary from real local traffic, precisely so it's more capable when it's cut off. Map tiles, geocoding, and hazard data are all cached locally with that scenario in mind.

The picture in my head has always been: grid's down, the box is running on backup power serving wifi, and my neighbors connect with their phones and everyone can see what's happening around us, instead of one guy hunched over a scanner relaying it. It's not shelf gear you dust off when something happens; the whole point is that running it every day is what trains it for the day you need it. Transcription is already fully local, the LLM extraction step currently uses a hosted model when online (it runs me about a dollar a day at 24/7), and the next build phase is a GPU box that closes that last gap with a local model.

The feature people are having the most fun with is the airband/ADS-B correlation. The system pulls spoken callsigns out of airband transcripts and matches them to ADS-B tracks by time and geometry, with match windows scaled by altitude and speed, plus hex-to-tail-number derivation. Click a plane on the map and hear what its pilot actually said, with a confidence badge when the match came from a partial readback. Beyond that: an ATAK feed (CoT/KML/GeoJSON), twice-daily AI briefs, a BOLO board, incident threading with a unit activity board, and helicopter orbit detection.

ADS-B aircraft position data automatically linked to airband voice transcripts
ADS-B aircraft position data automatically linked to airband voice transcripts

Benjamin adds that a public demo is available at https://demo.scannercommand.com. But he notes that "the public demo is a replay of two real days of traffic from my setup".

The software has not yet been released, and it does not appear to be free or open-source. There is a signup waitlist available on the scannercommand.com website.

This is an exciting time with AI not only helping to rapidly develop new software, but now being used to summarize and condense the vast amount of information about the state of the surrounding world available in the RF spectrum. 

Scammer Command example mapped incident report
Scammer Command example mapped incident report
I Turned My Police Scanner Into a Live AI Incident Map (24/7)

Wavelingo: Real-Time AI Shortwave Translation for the KiwiSDR Network

Thank you to Christophe (F4DAN) for writing in about his new project called Wavelingo, an AI real-time shortwave radio translator. The software currently works with the KiwiSDR web SDR network. Christophe has a live public example running at wavelingo.app, however, with a 60-second timeout due to hosting cost constraints. Christophe writes:

Are you listening to a QSO in a foreign language on your transceiver? Click on the closest SDR (KiwiSDR fleet for now, more SDR to come in the future), and get real-time translations.

I opened a telegram channel to share updates and feedbacks on this projects - and provide support.

The software is open source, with the code available on GitHub.

Wavelingo - Real time AI translationfor shortwave radio stations received via the KiwiSDR network.
Wavelingo - Real-time AI translation for shortwave radio stations received via the KiwiSDR network.

Automatic Signal Recognition with AI Machine Learning and RTL-SDR

Thank you to Trevor Unland for submitting his AI machine learning project called "RTL-ML" which automatically recognizes and classifies eight different signal types on low-power ARM processors running an RTL-SDR.

Trevor's blog post explains the machine learning architecture in detail, the accuracy he obtained, and how to try it yourself. If you try it for yourself, you can either run the pre-trained model or train your own model if you have sufficient training data.

The code is entirely open source on GitHub, and the training set data has been shared on HuggingFace. 

RTL-ML is an open-source Python toolkit for automatic radio signal classification using machine learning. It runs on ARM single-board computers like the Raspberry Pi 5 or Indiedroid Nova paired with an RTL-SDR Blog V4, achieving 87.5% accuracy across 8 real-world signal types including ADS-B aircraft transponders, NOAA weather satellites, ISM sensors, FM broadcast, NOAA weather radio, pagers, and APRS.

The project provides a complete pipeline from signal capture to trained classifier. Unlike academic approaches that rely on synthetic data or expensive GPU hardware, RTL-ML uses real signals captured from actual antennas and runs entirely on edge hardware with no cloud dependency. The Random Forest model is 186KB and processes signals in around 120ms on a Pi 5.

The GitHub repository includes the full capture and training scripts, a pre-trained model, 8 validated spectrograms, and documentation for adding new signal types. It works out of the box on both Raspberry Pi 5 and Indiedroid Nova with identical code and accuracy.

RTL-ML Setup: RTL-SDR Blog V4, Dipole Antenna and Indiedroid Nova ARM Computer.
RTL-ML Setup: RTL-SDR Blog V4, Dipole Antenna and Indiedroid Nova ARM Computer.

You might also be interested in some similar projects we've posted about in the past, such as this Shazam-style signal classifier, which used audio data from sigidwiki.com, and an Android app doing the same thing (which unfortunately now appears to have been removed from Google Play). There is also this deep learning based signal classifier model.

GhostHunter (Anti-LIF): Using Spiking Neural Networks to Rescue Satellite Signals Drowned in Noise

Thank you to Edwin Temporal for writing in and showing how his proprietary neuromorphic engine, GhostHunter (Anti-LIF), is being used to recover satellite data buried in the noise floor, which typical DSP methods would fail to do.

To recover the signals, Edwin uses trained Spiking Neural Networks (SNN). SNNs are artificial neural networks that draw further inspiration from nature by incorporating the 'spiking' on/off behavior of real neurons. Edwin writes:

My engine has successfully extracted and decoded structured data from high-complexity targets by mimicking biological signal processing:

Technosat: Successful decoding of GFSK modulations under extreme frequency drift and low SNR conditions.

MIT RF-Challenge: Advanced recovery of QPSK signals where traditional digital signal processing (DSP) often fails to maintain synchronization.

These missions are fully documented in the https://temporaledwin58-creator.github.io/ghosthunter-database/, which serves as a public ledger for my signal recovery operations. Furthermore, the underlying Anti-LIF architecture is academically backed by my publication on TechRxiv, proving its efficiency in processing signals buried deep within the noise floor.

Although the engine remains proprietary, I provide comprehensive statistical reports and validation metrics for each mission. I believe your audience would be thrilled to see how Neuromorphic AI (SNN) is solving real-world SIGINT challenges.

In the database, Edwin shows how his Anti-LIF system has recovered CW Morse code telemetry and QPSK data from noisy satellite signals. 

While Edwin's Anti-LIF is proprietary, he is offering proof of concept decoding. If you have a 250MB or less IQ/SigMF/Wav recording of a signal that is buried in the noise floor, you can submit it to him via his website, and he will run Anti-LIF on it for analysis.

Advanced readers interested in AI/neural network techniques for signal recovery can also check out his white paper on TechRxiv, where he shows signal recovery from signals buried in WiFi noise, as well as results from use in ECG and Healthcare applications.

An Example Signal Recovery with the Anti-LIF Spiking Neural Network
An Example Signal Recovery with the Anti-LIF Spiking Neural Network