Playlist Engine
A new mix every hour based on energy curve, BPM and Camelot key. No manual selection.
Your radio station.
24/7 live.
SAITS HyperRadio automates playlist generation, AI voice-overs, news blocks, jingles and monitoring.
2 stations live99.9% uptime24/7 since July 2026
Now playing
Losing It
Fisher · Catch & Release
Queue · autopilot
- 14:07Jingle — SplashFM ID
- 14:09Fisher — Losing It
- 14:15Overlay line — djgerard
- 14:18CamelPhat — Cola
AI Voice
News, weather, jingles and DJ talkovers generated in real time. Dutch and English.
Live Studio
Grab the mic in your browser. Sidechain ducking, 0.5s crossfade, no noticeable latency.
Multi-Tenant
Own database, pipeline, admin and stream per station. Fully separated.
Self-Healing
Watchdog, stream monitor and alerts. Fault detected? Automatic restart.
Why
Running radio 24/7 takes a lot of manual work
Building playlists, cutting jingles, reading the news, monitoring the stream and fixing problems — normally jobs for several people.
Data flow
From track to listener in 6 steps
- Track upload — the analyzer measures BPM, key, energy and cue points
- Playlist generation based on energy curve and musical compatibility
- Storage in DynamoDB with an hour key (Europe/Amsterdam, DST-aware)
- The pusher feeds tracks into the Liquidsoap queue at the right moment
- Liquidsoap plays with crossfade, cue points and sidechain ducking
- Icecast streams live in MP3, AAC and Opus
Building playlists by hand is a chore
A new mix every hour that fits in energy, BPM and key. Our generator does this automatically based on audio analysis of every track.
Stream outages cost you listeners
The self-healing pipeline detects outages and recovers automatically.
Recording voice-overs takes time
The AI voice engine generates jingles, news and sweepers in real time, in your own voice.
How it works
Three steps to your own station
From a folder of tracks to a live stream.
Upload your music
Drag your tracks into the library. The analyzer automatically measures BPM, key, energy and cue points for every track.
Set up your format
Choose your music style, jingles, news blocks and voice-over personality. Define your hourly schedule — how many tracks, when news, when jingles.
Go live
Your station starts immediately and runs 24/7. Listeners tune in via your own URL. You get a dashboard with real-time stats, now-playing info and monitoring.
Features
Everything a radio station needs in one platform
24/7 Automation
Playlists are generated every hour based on energy curve, BPM and harmonic mixing (Camelot keys). Jingles rotate automatically. News blocks on the hour, cut mix on the half hour. All driven by timers.
AI Voice Engine
ElevenLabs voice synthesis with multiple voice personalities. News, weather, jingles, sweepers and DJ talkovers — generated in real time. Dutch and English.
Multi-Tenant
One server, multiple stations. Every station has its own database, audio pipeline, admin portal and stream. Fully separated.
Live DJ Takeover
Open the Live Studio in your browser, click “Go Live” and your voice goes straight over the music — with sidechain ducking and a 0.5s crossfade.
Track Analysis
Every track is analysed with librosa: BPM, Camelot key, energy score, brightness, bass energy, cue points, silence detection and peak/RMS levels. The generator builds sets that make musical sense — not random.
Monitoring & Self-Healing
Stream monitor, watchdog, alert monitor and audio quality logger run 24/7. On failure: automatic restart of containers or services. Slack notifications for everything you need to know.
The Studio
This is what your control room looks like
Real screenshots from the SAITS Live Studio. No mockups: this is the environment DJs and newsrooms work in every day.
01 · Live desk
Your complete control room, in the browser
Three players with waveform, BPM and energy, an hour log that follows along and a multi-track transition editor. Everything you expect from a turntable setup, with automation around it.
- Players with cue, loop and PFL
- Transitions to the millisecond
- Music, jingles, news and voice on their own track
02 · Monitoring
Always know what happens in the next hours
The playlist horizon shows four hours ahead, the queue shows what is coming up and the status row flags immediately when the stream, sources or audio deviate.
- Playlist horizon per hour
- Live queue with jingles and music
- Broadcast status with alarms and sources
03 · Transition editor
Mix, cut and voice-over to the millisecond
Multiple tracks for music, jingles, news and voice. Cut, split and fade directly on the waveform, record a voice-over and adjust ducking and gain per clip.
- Tracks per deck, news and voice
- Cut, split, fade and snap on the waveform
- Record voice-overs with ducking and gain
04 · Modular
Every component can also be opened on its own
All modules run standalone: an hour clock on a second screen, a queue on the tablet, the jingle machine on the mixing desk. Presets bundle them per role.
- 40+ standalone studio objects
- Presets for DJ, nonstop, live and jingles
- Status light with real IoT status
05 · Your look
Pick the mood that fits your station
From Ocean to Sunset and a classic playout look: every theme restyles the whole studio. Handy when several stations or DJs share the same environment.
- 20+ colour themes
- Per station or per user
- Classic playout style available
Use cases
Who it is built for
Hospitality & Retail
One station for all your locations. Your own jingles, a consistent atmosphere, and optional video screens via HLS.
Brand Radio
Your own station with your name, voice and content — 24/7 on your site, with live video.
Community & Web Radio
Upload your tracks and your station is live on the cluster. Icecast, Liquidsoap and Kubernetes are configured automatically.
Existing broadcasters
From manual playout to full automation on Kubernetes. Your volunteers focus on content, not on technology.
Live video + HLS
Radio with pictures: 24/7 scheduled video playout, break blocks, live DJ sets and HLS output. Your own TV channel next to your radio station.
Kubernetes 24/7
Runs on EKS Anywhere Kubernetes with Rancher Fleet GitOps. Self-healing pods, Prometheus/Grafana monitoring and cert-manager TLS.
Architecture
How it is built
One physical server with Docker-containerised audio pipelines. Every station has its own Liquidsoap engine, Icecast server and database. On-premise, no cloud dependency.
Liquidsoap 2.4.0Icecast 2.5.0DynamoDB LocalPython FlaskWeb Audio APIWebSocketElevenLabs APIlibrosaEssentiaKubernetes (EKS Anywhere)Rancher FleetGitOps
Book a demo
Start your radio station
You decide the music and the style — SAITS HyperRadio automates the rest.
- Reply within one working day
- No obligation
- We only use your details for this request