Background Music HQ Generator

Generate production-grade background music for presentations, podcast jingles, keynotes, branded per-pet social media tracks, and professional content using MiniMax Music 2.5 on Replicate.

This skill is the high-quality counterpart to bg-music. While bg-music uses Stable Audio 2.5 for quick, lightweight tracks (up to 90s), this skill uses MiniMax Music 2.5 which produces longer, richer, studio-quality tracks with proper song structure, style-aware mixing, and support for 100+ instruments.

It includes 11 pet-specific mood presets designed for social media content — from playful zoomie reels to heartfelt adoption stories to epic slow-mo showcases — making it ideal for pre-generating branded background music per pet.

When to Use This vs bg-music

Setup

Install dependencies (one-time):

pip3 install -r ~/.cursor/skills/bg-music-hq/scripts/requirements.txt

Set your Replicate API token (skipped if already configured by another skill):

python3 ~/.cursor/skills/bg-music-hq/scripts/setup_key.py YOUR_REPLICATE_API_TOKEN

Or auto-import from an existing skill:

python3 ~/.cursor/skills/bg-music-hq/scripts/setup_key.py

Quick Reference

Generate presentation background music

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "clean modern background for tech startup pitch" \
  --mood presentation --output presentation_bgm.mp3

Generate a podcast intro jingle

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "catchy upbeat podcast opener" \
  --mood podcast-intro --output podcast_intro.mp3

Generate a keynote backing track

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "inspiring cinematic build for product launch keynote" \
  --mood keynote --output keynote_music.mp3

Generate with custom lyrics and vocals

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  --lyrics lyrics.txt \
  --prompt "Indie folk, warm, acoustic guitar, male vocals" \
  --with-vocals --output song.mp3

Generate high-quality WAV for production

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "orchestral background for documentary" \
  --mood cinematic --format wav --output documentary_score.wav

Generate variations

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "smooth jazz for podcast bed" \
  --mood podcast-bed --variations 3 --output podcast_bed.mp3
# Generates: podcast_bed_1.mp3, podcast_bed_2.mp3, podcast_bed_3.mp3

generate_bgm_hq.py Options

Mood Presets

Presentation & Podcast Moods

Pet Content & Social Media Moods

General Moods

Moods are defined in scripts/moods.json and can be customized.

Structure Tags

MiniMax Music 2.5 supports 14 structure tags for precise arrangement control. The mood presets include pre-built structure templates, but you can override them with --lyrics or --lyrics-text:

For instrumental background music (the default), the script uses [Inst] tags with parenthetical instrument directions.

Common Workflows

Per-Pet Branded Music (Mascotify)

Pre-generate a branded music library for each pet, matching their personality to a mood:

# Playful golden retriever — bouncy, fun tracks for zoomie reels
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "bouncy cheerful background for a playful golden retriever puppy" \
  -m pet-playful -d 20 -n 2 -o assets/bgm/pets/max_playful.mp3

# Regal Persian cat — elegant, sophisticated ambiance
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "elegant classical-inspired background for a majestic Persian cat" \
  -m pet-regal -d 25 -o assets/bgm/pets/luna_regal.mp3

# Chill old bulldog — cozy lazy-afternoon vibes
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "warm cozy lo-fi background for a sleepy old bulldog" \
  -m pet-chill -d 30 -o assets/bgm/pets/bruno_chill.mp3

# Energetic border collie — adventure/outdoor tracks
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "driving adventurous background for an energetic border collie on hikes" \
  -m pet-adventure -d 25 -o assets/bgm/pets/rocky_adventure.mp3

Pet Reel Music Library

Generate a complete set of tracks for different reel types:

# Goofy fail compilation
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "funny quirky background for pet fail compilation" \
  -m pet-goofy -d 15 -o assets/bgm/reels/goofy_fails.mp3

# Grooming transformation / rescue glow-up
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "dramatic build with uplifting reveal for pet grooming transformation" \
  -m pet-transformation -d 20 -o assets/bgm/reels/glow_up.mp3

# Heartfelt adoption story
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "emotional acoustic background for rescue dog adoption journey" \
  -m pet-heartfelt -d 30 -o assets/bgm/reels/adoption_story.mp3

# Trendy viral reel format
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "catchy modern beat for trending pet reel format" \
  -m pet-trendy -d 20 -o assets/bgm/reels/viral_trend.mp3

# Epic slow-mo showcase
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "epic cinematic track for slow-motion dog running on beach" \
  -m pet-epic -d 25 -o assets/bgm/reels/epic_slowmo.mp3

# Day-in-the-life montage
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "warm easygoing track for daily routine pet montage" \
  -m pet-daily -d 25 -o assets/bgm/reels/daily_life.mp3

# Sleeping puppy / bedtime content
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "ultra-soft lullaby for sleeping puppy ASMR content" \
  -m pet-lullaby -d 30 -o assets/bgm/reels/puppy_lullaby.mp3

Presentation backing track

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "clean modern background for a SaaS product demo" \
  -m presentation -o assets/bgm/demo_presentation.mp3

Podcast full set (intro + bed + outro)

# Intro jingle
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "energetic tech podcast opener" \
  -m podcast-intro -o assets/bgm/podcast_intro.mp3

# Interview bed
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "subtle background for interview conversation" \
  -m podcast-bed -o assets/bgm/podcast_bed.mp3

# Outro
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "warm friendly podcast closing" \
  -m podcast-outro -o assets/bgm/podcast_outro.mp3

Keynote with cinematic build

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "epic cinematic build for startup launch event" \
  -m keynote --bpm 105 --key "D major" -o assets/bgm/launch_keynote.mp3

Custom lyrics with structure

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  --lyrics-text "[Intro]
(Atmospheric synth opening)

[Inst]
(Full arrangement, driving modern beat, synth bass)

[Build Up]
(Rising intensity, layered synths)

[Inst]
(Peak energy, all instruments)

[Outro]
(Gentle fade, reverb tail)" \
  --prompt "Modern electronic, tech-forward, crisp production, 110 BPM" \
  -o assets/bgm/custom_track.mp3

High-quality WAV for video production

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "orchestral score for documentary intro" \
  -m cinematic -f wav --sample-rate 44100 --bitrate 256000 \
  -o assets/bgm/documentary_score.wav

Try different styles for the same concept

python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "warm acoustic background for workshop" \
  -m workshop -n 3 -o assets/bgm/workshop.mp3
# Each variation is unique — the model produces different arrangements each time

Fade-In / Fade-Out

Background music tracks benefit from smooth fades. By default:

• Fade-in: 500ms gentle ramp
• Fade-out: 2000ms smooth tail

Use --fade-in and --fade-out to customize (in ms), or --no-fade to disable.

Fades are applied via ffmpeg after generation.

Duration Control

MiniMax Music 2.5 generates variable-length tracks (typically 2:30 to 4:30). Use --duration to trim the output to a maximum length:

# Trim to 30 seconds for a short jingle
python3 ~/.cursor/skills/bg-music-hq/scripts/generate_bgm_hq.py \
  "upbeat podcast intro" -m podcast-intro -d 15 -o jingle.mp3

Trimming happens before fades are applied, so the fade-out is always clean.

Model

Uses MiniMax Music 2.5 (minimax/music-2.5) on Replicate.

Prompt Tips

The prompt field steers the overall sound. A good prompt follows this pattern:

[Genre], [Mood/Emotion], [Tempo], [Key instruments], [Production style]

Examples:

• Corporate pop, clean, modern, 100 BPM, piano and subtle synths, wide soundstage
• Smooth jazz, sophisticated, relaxed 85 BPM, walking bass, brushed drums, piano
• Cinematic orchestral, inspiring, building, sweeping strings, brass, timpani
• Lo-fi hip-hop, chill, vinyl texture, warm midrange, 75 BPM
• Electronic, tech-forward, crisp transients, synth bass, 115 BPM

Configuration

Config at ~/.cursor/skills/bg-music-hq/config.json:

{
  "replicate_api_token": "YOUR_TOKEN",
  "default_mood": "presentation",
  "default_sample_rate": 44100,
  "default_bitrate": 256000,
  "default_format": "mp3"
}

Falls back to tokens from bg-music, sound-effects, avatar-video-reel, or brand-asset-studio skills.