Mcaster1Studio — AI Integration

Features & Status — v0.4.0-beta — Last updated: 2026-03-18

Overview

Mcaster1Studio integrates AI capabilities across the broadcast workflow — from music recommendations and playlist generation to artist research and on-air persona management. The AI system is provider-agnostic, supporting six backends, and operates through a persona framework that tailors AI behavior to the station’s format, time of day, and surface context.

~35%AI Features Complete

6AI Providers

15Preset Personas

8Research Buttons

5SQL Dialects

Status: Core infrastructure (provider abstraction, persona system, artist intel, category AI tools) is functional and in active testing/QA. Streaming token support, advanced AuxDeck routing, and some animated UI elements are upcoming.

AI Providers

Configure your preferred provider and API key in Preferences > AI. All providers use a common request/response abstraction — switching providers does not affect feature behavior.

Provider	Type	Notes	Status
Ollama	Local / self-hosted	No API key required; runs on local hardware	Testing
Claude	Cloud (Anthropic)	API key required; supports Claude 3+ models	Testing
ChatGPT	Cloud (OpenAI)	API key required; supports GPT-4o and later	Testing
Grok	Cloud (xAI)	API key required	Testing
Gemini	Cloud (Google)	API key required; supports Gemini 1.5+	Testing
Venice	Cloud (Venice.ai)	API key required; privacy-focused provider	Testing

AI Persona System

15 preset AI personas ship with Mcaster1Studio, covering the full range of broadcast formats. Personas use a 3-tier resolution model: daypart schedule > surface-level persona > global default. Each persona has a customizable system prompt (500 char max) and can be assigned per-category via right-click.

Radio DJ Personas (7) Done

Top 40 / CHR — High energy, trend-focused
Classic Rock — Storytelling, deep cuts
Country — Warm, community-focused
Urban / Hip-Hop — Culture-forward, slang-aware
Jazz / Blues — Sophisticated, smooth delivery
EDM / Dance — Festival energy, BPM-aware
Adult Contemporary — Polished, relatable

Podcast Personas (3) Done

Interview Host — Conversational, guest-focused
Solo Narrator — Storytelling, pacing-aware
Panel Moderator — Multi-voice balance

Specialty Personas (5) Done

Sports Broadcaster — Stats-driven, play-by-play
TV News Anchor — Authoritative, concise
Church / Worship — Reverent, scripture-aware
Social Media Host — Casual, engagement-focused
Producer — Technical, workflow-oriented

Persona Infrastructure

Component	Detail
PersonaManager	Singleton with 3-tier resolution (daypart > surface > global) and caching
Database tables	`ai_personas` + `daypart_schedule` in all 5 SQL dialects
Preferences > AI	Persona combo selector + editable system prompt (500 char max)
Per-category assignment	Right-click any library category to assign a governing persona

Artist Intel Dialog

A full-window research dialog accessible via right-click on any track in the Media Library. Combines AI-generated research with manual notes, persistent per-tab JSON storage, and direct integration with the library track list via AI Intel badges.

Tabs Done

Overview — Summary bio, genre, active years, notable works
Discography — Albums, singles, compilations with release dates
Images — Artist photos and album artwork (cache-backed)
DJ Script — Ready-to-read on-air talking points

8 AI Research Buttons Done

Touring History — Concert tours, notable venues
Musical Influences — Influence tree (in/out)
Awards & Charts — Grammys, Billboard, certifications
Broadcaster Script — Pre-written DJ patter
Fan Base & Impact — Cultural significance, demographics
Full Timeline — Chronological career events
Gear & Equipment — Instruments, studios, tools
Discovery — Early career breaks, origin story

Interaction Features

Multi-turn “Ask AI” chat with full conversation context maintained
Save/load per-tab as JSON — overview and research tabs persist independently
Right-click tab context menu: Refresh Report, Save Individual Report as HTML
Auto-retry on busy — 5-second delay with user-friendly error messages
Save confirmation dialog showing database path, table name, and record ID
AI Intel badges appear in the library track list (ColIntel column) after saving research

AI DJ Agent Browser

An interactive conversational recommendation engine accessible from the Media Library’s category view.

Capabilities Testing

Persona-aware — recommendations reflect the active AI persona’s style and format
Conversational interface — multi-turn chat to refine recommendations
Category-scoped — draws from the active library category
Drag-to-deck — recommended tracks can be dragged directly to Deck A/B or the playlist queue
Animated braille spinner progress indicator with phase messages during AI processing

Playlist Generator Pro

AI-powered playlist generation with fine-grained control over output, broadcast element insertion, and persona-influenced track selection.

Capabilities Testing

Source selector — choose which library categories to draw from
AI Intel prioritization — tracks with saved AI research are weighted higher
Broadcast element insertion — station IDs, sweepers, jingles, and ads at configurable intervals
Duration targeting — generate playlists for specific time blocks (1h, 2h, 4h, custom)
Persona influence — the active persona shapes track selection, ordering, and element placement
Animated braille spinner with phase messages (Analyzing library, Selecting tracks, Arranging flow, Inserting elements)

Daypart Scheduler Pro

A 24-hour visual timeline for scheduling AI personas across the broadcast day. Each surface can have its own daypart schedule, persisted in the daypart_schedule table.

Capabilities Testing

24h timeline — visual hour-by-hour grid showing which persona is active
Persona per time block — assign different personas to morning drive, midday, afternoon, evening, overnight
AI auto-schedule — let the AI suggest a daypart layout based on station format and target audience
Per-surface scheduling — each surface can have its own daypart schedule
Database-backed — schedules persist across all 5 SQL backends
Drag-to-resize time blocks in the visual editor

Media Library AI Features

AI capabilities integrated directly into the Media Library module, enhancing search, categorization, and track discovery.

Search & Discovery Done

FTS5 full-text search — SQLite FTS5 virtual table for fast search; LIKE fallback if FTS5 unavailable
AI Intel badges — tracks with saved research display badge in ColIntel column
FTS5 search scoped to category — restricts results to active category
Album art cache — memory + disk cache with TagLib extraction

Category Management Done

7 preset categories — Music, Stingers, Station IDs, Sweepers, Jingles, Ads, Spoken Word
Hierarchical tree — 3+ levels, expandable/collapsible
Per-category persona — right-click to assign AI persona
Scan-into-category — auto-assign tracks based on path or metadata
7 new database tables + schema migration for existing databases

Upcoming

Features planned for near-term development. These are in design or early implementation.

Feature	Description	Status
AuxDeck 4-bus routing	Route AuxDeck output to 4 independent buses (AIR, CUE, REC, MON) for advanced monitoring workflows	Planned
Animated progress indicators	Expand braille spinner animations to all long-running AI operations across the application	Planned
Streaming token support	Streaming/chunked AI responses with real-time token display, reducing perceived latency	Planned
AI-powered show prep	Automated pre-show research packets combining artist intel, trending topics, and daypart context	Planned
Voice cloning integration	AI voice synthesis for pre-recorded liners and sweepers using station voice talent samples	Planned