open-source-music-generation

open-source-music-generation is a skill for Claude Code, Codex from S3YED/appie-kit. It costs 53 tokens per session (2,155 once invoked), scanned A, original, MIT.

A guide to open-source tools that generate music or sound from text, including tools for songs with lyrics and tools for instrumental music or sound effects.

In plain words
What is it for?
It helps generate lyric-based songs, text-to-music tracks, text-to-sound effects, melody-conditioned music, and stereo audio.
Why use it?
It explains which tool fits different audio tasks and covers setup, model choices, generation settings, and common problems.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit It helps generate lyric-based songs, text-to-music tracks, text-to-sound effects, melody-conditioned music, and stereo audio.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/s3yed/appie-kit/open-source-music-generation
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add S3YED/appie-kit --skill open-source-music-generation
Clone the repo
git clone --depth 1 https://github.com/S3YED/appie-kit

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for open-source-music-generation

README.md
[![agentmods](https://agentmods.dev/badge/skills/s3yed/appie-kit/open-source-music-generation/github.svg)](https://agentmods.dev/skills/s3yed/appie-kit/open-source-music-generation)
Your own site
<a href="https://agentmods.dev/skills/s3yed/appie-kit/open-source-music-generation"><img src="https://agentmods.dev/badge/skills/s3yed/appie-kit/open-source-music-generation/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for open-source-music-generation

Your own site · 80×15
<a href="https://agentmods.dev/skills/s3yed/appie-kit/open-source-music-generation"><img src="https://agentmods.dev/badge/skills/s3yed/appie-kit/open-source-music-generation.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 53 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,155 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00053 $0.02155
Opus 5 $0.00026 $0.01077
Sonnet 5 $0.00011 $0.00431
Haiku 4.5 $0.00005 $0.00215

Measured 5d ago against content hash 0af2756395cc, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

open-source-music-generation scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/content/open-source-music-generation/SKILL.md · 214 lines

How it starts

The opening of the file, as written. The whole thing — 214 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Open-Source Music Generation

A unified guide to two major open-source music generation ecosystems: HeartMuLa and AudioCraft (MusicGen/AudioGen). Both generate audio from text descriptions; HeartMuLa specializes in lyrics-conditioned song generation while AudioCraft offers broader text-to-music and text-to-sound capabilities with melody conditioning.

Quick Comparison

Feature HeartMuLa AudioCraft
Focus Lyrics+tags song generation Text-to-music, text-to-sound
Model sizes 3B, 7B 300M (small) to 3.3B (large)
Output MP3 48kHz stereo WAV 32kHz mono/stereo
VRAM (min) ~6.2GB (3B lazy) ~2GB (small fp16) to ~16GB (large)
Lyrics support Core feature (required) Not supported (text-only)
Melody conditioning No Yes (musicgen-melody)
License Apache-2.0 MIT
Setup complexity Medium (patches needed) Low (pip install)

When to use HeartMuLa: You need to generate a full song with lyrics, multiple languages, or want the most open-source-capable Suno alternative.

When to use AudioCraft: You need text-to-music or text-to-sound effects, melody conditioning, stereo output, or style transfer.


HeartMuLa — Lyrics-to-Song Generation

HeartMuLa generates full songs from lyrics + genre/style tags. Families of models: HeartMuLa (music language model), HeartCodec (12.5Hz audio codec), HeartTranscriptor (lyrics transcription), HeartCLAP (audio-text alignment).

Hardware Requirements

  • Minimum: 8GB VRAM with --lazy_load true (loads/unloads models sequentially, peaks ~6.2GB)
  • Recommended: 16GB+ VRAM for single-GPU
  • Multi-GPU: --mula_device cuda:0 --codec_device cuda:1
  • CPU mode: Possible with --mula_device cpu --codec_device cpu but extremely slow (30-60 min per song)

Installation

# 1. Clone
git clone https://github.com/HeartMuLa/heartlib.git
cd heartlib

# 2. Python 3.10 venv
uv venv --python 3.10 .venv
. .venv/bin/activate
uv pip install -e .

# 3. Fix dependency conflicts
uv pip install --upgrade datasets transformers

# 4. Download models (all in parallel)
hf download --local-dir './ckpt' 'HeartMuLa/HeartMuLaGen'
hf download --local-dir './ckpt/HeartMuLa-oss-3B' 'HeartMuLa/HeartMuLa-oss-3B-happy-new-year'
hf download --local-dir './ckpt/HeartCodec-oss' 'HeartMuLa/HeartCodec-oss-20260123'

Read the full file on GitHub · 214 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 214 lines · 53 tokens per session scan A 0af2756395cc

Subscribe to this mod's changes

open-source-music-generation is a skill published in the GitHub repository S3YED/appie-kit (8 stars, last pushed 13d ago), licensed MIT. It adds 53 tokens to every session and 2,155 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.