live-radio/README.md

Live Ozan Radio

Personal AI radio — no catalog music, ever. DeepSeek is the DJ. Google Lyria 3 composes every track. Spotify is read-only taste input.

Inspired by Magenta RealTime 2 (live, ~200ms) and Lyria 3 (full songs via Gemini API). On Mac you can layer MRT2 for true live improvisation; this repo ships the cross-platform Lyria + DeepSeek stack first.

Stack

Layer	Product	Role
DJ brain	DeepSeek (via Tinqs inference or BYOK)	Mood, prompts, variety
Music engine	Google Lyria 3 Pro / Clip	Generate MP3 tracks
Taste	Spotify Web API (optional)	Top artists, genres — never plays Spotify
Player	FastAPI + gateway/player.html	Stream generated queue
Live (optional)	Magenta RealTime 2	Apple Silicon only — see below

Taste (settings.json)

Edit settings.json at the repo root — the DJ reads it on every generate and chat. Default profile: ethnic world dubtronica (global roots + dub space + electronic groove).

Saved songs

Every track is written to ./songs/ and committed via Git LFS (audio) + plain git (metadata):

File	Storage	Contents
{id}_{title}.mp3	LFS	Audio
{id}.meta.json	git	Title, mood, DJ line, prompt, lyrics, timestamp

Browse in the player under Saved songs, or GET /api/songs. After clone: git lfs install then git lfs pull.

Quick start (Forge / Windows)

cd live-ozan-radio
python -m venv .venv
.venv\Scripts\activate
pip install -e .
copy .env.example .env
# Fill GEMINI_API_KEY + DEEPSEEK_API_KEY (and Spotify if you have them)

python -m ozan_radio serve
# Open http://127.0.0.1:8787/player

One-shot track (no server):

python -m ozan_radio generate

Environment

Variable	Required	Notes
GEMINI_API_KEY	Yes	Google AI Studio — Lyria 3
DEEPSEEK_API_KEY	Yes	Tinqs proxy or DeepSeek direct
DEEPSEEK_BASE_URL	No	Default https://tinqs.com/api/v1/inference
SPOTIFY_*	No	Refresh token flow — taste only
LYRIA_MODEL	No	lyria-3-pro-preview (default) or lyria-3-clip-preview

Spotify setup (taste profile)

1. Create an app at Spotify Developer Dashboard.
2. Add redirect URI http://127.0.0.1:8888/callback.
3. Complete OAuth once to obtain a refresh token (scope: user-top-read).
4. Paste SPOTIFY_CLIENT_ID, SPOTIFY_CLIENT_SECRET, SPOTIFY_REFRESH_TOKEN into .env.

The Spotify MCP / tool you wired to DeepSeek can call the same endpoints — this repo exposes them natively for the DJ loop.

API

Method	Path	Description
GET	/api/now	Current track metadata
GET	/api/queue	Full queue
POST	/api/generate	DJ plans + Lyria renders next track
POST	/api/skip	Advance queue
GET	/stream/{file}	MP3 stream
GET	/player	Web UI

Magenta RealTime 2 (optional live layer)

On Apple Silicon (Kraken), install magenta-rt for sub-second live generation:

uv pip install "magenta-rt[mlx]"
mrt models init && mrt models download
mrt mlx generate --prompt "disco funk" --duration 4.0 --model=mrt2_small

Wire MRT2 as a bridge between tracks or as a live “bed” under the Lyria queue — PRs welcome.

Publish on tinqs.com (public repo)

1. On Git Studio: + → New Repository
   • Owner: tinqs
   • Name: live-radio
   • Visibility: Public
2. Push (with LFS for song MP3s):

git lfs install
git init
git remote add origin git@ssh.tinqs.com:tinqs/live-radio.git
git add .
git commit -m "Live Ozan Radio — DeepSeek DJ + Lyria 3"
git push -u origin main

3. Preview the player: https://tinqs.com/tinqs/live-radio/src/branch/main/gateway/player.html (static shell; audio streams from your running server).

Agent usage

DeepSeek (Pi, Cursor, Claude Code) can operate the station via HTTP or the skill in .cursor/skills/ozan-radio/SKILL.md.

License

Apache 2.0 — same spirit as Magenta RealTime 2 open weights.