What music does the AI play, and where does it come from?

The AI plays only from your own music library, pulled from Navidrome over the Subsonic API. It does not generate music, replace your taste, or draw from an external catalogue.

Can listeners skip songs or control what plays next?

No. Listeners cannot skip; track-end is the only natural transition. The DJ (AI plus operator-curated personas) owns the pacing. Operators can skip via the admin API, but listeners cannot.

What LLM providers and text-to-speech options are supported?

The LLM provider is swappable at runtime among Ollama, Anthropic, OpenAI, Google, DeepSeek, OpenRouter, Vercel AI Gateway, or any OpenAI-compatible server. Five TTS engines are available: Piper and Kokoro in-process, plus optional Chatterbox (zero-shot voice cloning) and PocketTTS (6× real-time, EN/FR/DE/IT/ES/PT) in a sidecar, and cloud options from OpenAI or ElevenLabs.

How do listeners access the broadcast?

Listeners connect via a single shared Icecast stream using native iOS or Android apps (available on the App Store and Google Play), a web player, or a PWA. The broadcast simultaneously serves MP3 128 kbps and Ogg-Opus 96 kbps so the player can pick the best codec.

Back to articlesLarge Language Models

Large Language Models

SUB/WAVE launches a self-hosted AI radio station that curates your music library and speaks between tracks like a human DJ.

Hacker News13h ago6 min read

Key takeaway

SUB/WAVE is a self-hosted radio station generator that uses an AI DJ to curate and broadcast tracks from your personal music library on a single shared stream, with support for multiple LLM providers and text-to-speech engines. It is designed for operators who want radio-style curation (one shared broadcast, no skip button for listeners) rather than playlist-based streaming, and it runs on a single Linux host behind optional Cloudflare.

Summaries like this, in your inbox every morning.

3 Key Points

What happened
SUB/WAVE is a self-hosted radio platform that streams a single shared Icecast broadcast where an AI DJ picks tracks from your music library (via Navidrome), writes intros, and reads station idents, weather, and time checks. Listeners can make song requests in plain language; the platform supports swappable LLM providers (Ollama, Anthropic, OpenAI, Google, DeepSeek, OpenRouter, Vercel AI Gateway) and five TTS engines, with native iOS and Android apps plus a web player.
Why it matters
Unlike playlists or on-demand streaming, SUB/WAVE creates a shared broadcast experience where everyone hears the same thing at the same time—radio rather than a jukebox. The operator (not the listener) controls pacing, and the AI pulls only from your own music library, so it curates based on your taste rather than external algorithms or generated content. For self-hosters and small broadcasters, this offers a way to automate station-like curation without external music licensing or cloud dependencies.
What to watch
Setup requires Docker, a Navidrome music server, and choice of LLM provider; the CLI installer or docker-compose method takes three configuration variables. The platform includes hourly MP3 archives, a visual "Library Observatory" map of tagged tracks by genre and energy, scheduled shows with up to 10 DJ personas, and pluggable skills (weather, news, traffic, custom). iOS and Android apps are available on their respective app stores.

FAQ

What music does the AI play, and where does it come from?: The AI plays only from your own music library, pulled from Navidrome over the Subsonic API. It does not generate music, replace your taste, or draw from an external catalogue.
Can listeners skip songs or control what plays next?: No. Listeners cannot skip; track-end is the only natural transition. The DJ (AI plus operator-curated personas) owns the pacing. Operators can skip via the admin API, but listeners cannot.
What LLM providers and text-to-speech options are supported?: The LLM provider is swappable at runtime among Ollama, Anthropic, OpenAI, Google, DeepSeek, OpenRouter, Vercel AI Gateway, or any OpenAI-compatible server. Five TTS engines are available: Piper and Kokoro in-process, plus optional Chatterbox (zero-shot voice cloning) and PocketTTS (6× real-time, EN/FR/DE/IT/ES/PT) in a sidecar, and cloud options from OpenAI or ElevenLabs.
How do listeners access the broadcast?: Listeners connect via a single shared Icecast stream using native iOS or Android apps (available on the App Store and Google Play), a web player, or a PWA. The broadcast simultaneously serves MP3 128 kbps and Ogg-Opus 96 kbps so the player can pick the best codec.

Discussion

No comments yet. Be the first to share your thoughts!

Unable to generate summary — the article body provided contains no news content, only a newsletter registration prompt.

Nikkei AI Stocks1h ago

GitHub Copilot's shared AI harness matches rival model-vendor tools on task completion while using fewer tokens, letting developers mix-and-match models without sacrificing performance.

GitHub Copilot Blog4h ago

OpenAI will release GPT-5.6 in limited preview with Trump administration case-by-case approval, a less restrictive arrangement than the export controls imposed on rival Anthropic.

The Verge AI4h ago

The Trump administration is pressuring OpenAI to limit early access to its new GPT 5.6 model to select partners before a broader public release, shifting from its previous hands-off AI stance.

TechCrunch AI4h ago

Visa partners with AI and stablecoin firms to build payments infrastructure for autonomous agents and digital assets, diversifying revenue beyond traditional card fees.

Top Companies AI — US (1/2)7h ago

TrueFoundry acquires MLOps pioneer Seldon AI to combine infrastructure for deploying AI agents at scale, addressing a gap where only 14% of enterprises have moved AI pilots into production.

Top Companies AI — US (1/2)7h ago

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime

5 minutes a day. The AI essentials.

200+ sources · Email / LINE / Slack

Get it free →