ACE-Step-Studio AGENTS.md

Repository guidance for ACE-Step Studio, a self-contained application that generates music with AI. It describes the bundled Python, Node.js, FFmpeg, model files, and the separate processes used to run the app.

In plain words
What is it for?
Use it when working on the launcher, development setup, music-generation pipeline, model switching, or the web and API processes.
Why use it?
It helps an agent understand a project that combines several runtimes and services, so changes can target the correct part.

Instructions file for CodexOpenCode

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.

agentmods
npx agentmods add instructions/timoncool/ace-step-studio/agents-md
Clone the repo
git clone --depth 1 https://github.com/timoncool/ACE-Step-Studio

Made for: Codex, OpenCode.

Per session 3,105 This file is loaded in full into every session.
When invoked 3,105 The same file — it is already loaded in full.
Security scan A 1 finding. Scan, not verified.
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 $0.03105 $0.03105
Opus 5 $0.01553 $0.01553
Sonnet 5 $0.00621 $0.00621
Haiku 4.5 $0.00311 $0.00311

Measured 2d ago against content hash 48ef4fe036d1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ACE-Step-Studio AGENTS.md scanned grade A with 1 finding 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 2d 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.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- Spawns Python with `child_process.spawn()`, pipes stdout/stderr with `[Gradio]` prefix
AGENTS.md · 223 lines

How it starts

The opening of the file, as written. The whole thing — 223 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ACE-Step Studio — Agent Guidelines

Project Overview

ACE-Step Studio is a portable, self-contained AI music generation application. It bundles Python (with CUDA/PyTorch), Node.js, and the ACE-Step 1.5 ML pipeline into a single-terminal launcher. No installation required — extract and run.

GitHub: https://github.com/timoncool/ACE-Step-Studio Upstream: Forked from fspecii/ace-step-ui, heavily modified.

Architecture

ACE-Step-Studio/
├── run.bat                    # Single-terminal launcher (prod)
├── run-dev.bat                # 3-terminal dev mode with Vite HMR
├── install.bat                # First-time setup
├── run-no-lm.bat              # Launch without LM (more VRAM)
├── python/                    # Portable Python 3.12 + CUDA 12.8
├── node/                      # Portable Node.js
├── ffmpeg/                    # Portable FFmpeg
├── models/                    # HuggingFace model cache
├── ACE-Step-1.5/              # ML pipeline (Gradio + FastAPI on port 8001)
│   ├── acestep/               # Python package
│   │   ├── acestep_v15_pipeline.py  # Main entry point
│   │   ├── llm_inference.py         # 5Hz LM (vLLM/nanovllm)
│   │   ├── api/http/                # FastAPI routes (/v1/init, /v1/models, /health)
│   │   └── ui/gradio/api/           # Gradio API routes
│   └── checkpoints/           # DiT + LM model weights
└── app/                       # Web application
    ├── index.html             # Entry point (uses CDN for Tailwind + esm.sh)
    ├── App.tsx                # Root component, generation logic
    ├── components/            # React components
    │   ├── CreatePanel.tsx    # Generation UI (Simple + Custom modes)
    │   ├── SongList.tsx       # Song list with model badges + generation time
    │   ├── RightSidebar.tsx   # Song details panel
    │   ├── Sidebar.tsx        # Left sidebar (GPU stats, navigation)
    │   └── ...
    ├── services/api.ts        # API client + Song data transformation
    ├── context/I18nContext.tsx # i18n provider
    ├── i18n/                  # Translations (en, ru, zh, ja, ko)
    ├── types.ts               # TypeScript interfaces
    ├── dist/                  # Vite production build (served by Express)
    └── server/                # Express backend (port 3001)
        ├── src/index.ts       # Server entry + static serving + pipeline manager
        ├── src/config/        # Config (ports, paths, pipeline settings)
        ├── src/routes/        # API routes (songs, generate, pipeline, etc.)
        ├── src/services/
        │   ├── pipeline-manager.ts  # Spawns/monitors/restarts Python process
        │   ├── acestep.ts           # Gradio client, job queue, generation logic
        │   ├── gradio-client.ts     # Gradio connection management
        │   └── storage/             # Local file storage
        └── src/db/            # SQLite migrations + connection pool

Read the full file on GitHub · 223 lines

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. 2d ago First seen · 223 lines · 3,105 tokens per session scan A 48ef4fe036d1

Subscribe to this mod's changes

ACE-Step-Studio AGENTS.md is an instructions file published in the GitHub repository timoncool/ACE-Step-Studio (327 stars, last pushed 9d ago), licensed MIT. It adds 3,105 tokens to every session, about $0.0155 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.