ACE-Step-Studio
llms.txt416 starsChanged 17 days agoMITView on GitHub
A Windows desktop studio for ACE-Step 1.5, the open music model whose language model plans a song and whose DiT renders it, running on the user's own NVIDIA card through the native C++ engine acestep.cpp. While it is open it serves an MCP server at http://127.0.0.1:8792/mcp that an agent uses to do everything the studio does: songs from a caption and lyrics, songs planned by ACE-Step's own language model, covers, remixes and repaints of a track, the library, covers, stems,…
We split this llms.txt into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Only agents on this computer and the studio's own window may connect.
≈ “Common” is an estimate: at least 1 in 5 of the 3,225 llms.txt files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
# ACE-Step Studio
> A Windows desktop studio for ACE-Step 1.5, the open music model whose language model plans a song and whose DiT renders it, running on the user's own NVIDIA card through the native C++ engine acestep.cpp. While it is open it serves an MCP server at http://127.0.0.1:8792/mcp that an agent uses to do everything the studio does: songs from a caption and lyrics, songs planned by ACE-Step's own language model, covers, remixes and repaints of a track, the library, covers, stems, MIDI from any track, karaoke, audio processing, video clips, the player, LoRA from a catalogue or Hugging Face, and a LoRA trained from a folder of songs end to end. The agent also sees the studio's window and works it like a user.
For an agent given this repository: install the studio from the latest release, start it, connect to its MCP server, and read the skill the server serves (resource `studio://skill`, prompt `studio`) - it names every tool, what the model reads, and the recipes. Start every session with the tool `studio_status`.
## Connect
- MCP endpoint: `http://127.0.0.1:8792/mcp` (Streamable HTTP; protocol 2026-07-28 stateless, and 2025-11-25 / 2025-06-18 / 2025-03-26 through `initialize`)
- Claude Code: `claude mcp add --transport http ace-step-studio http://127.0.0.1:8792/mcp`
- Other clients: `{ "mcpServers": { "ace-step-studio": { "type": "streamable-http", "url": "http://127.0.0.1:8792/mcp" } } }`
- Only agents on this computer and the studio's own window may connect; there is no key.
## Docs
- [Agent skill](https://raw.githubusercontent.com/timoncool/ACE-Step-Studio/main/docs/mcp-skill.md): every tool by area, the model's writing rules, step-by-step recipes (the same text the server serves as `studio://skill`)
- [README](https://github.com/timoncool/ACE-Step-Studio#readme): what the studio does, what it needs, models and where they go
- [Latest release](https://github.com/timoncool/ACE-Step-Studio/releases/latest): the installer and the portable archive
- [Project page](https://timoncool.github.io/ACE-Step-Studio/)
## Optional
- Install the skill for Claude Code: save [docs/mcp-skill.md](https://raw.githubusercontent.com/timoncool/ACE-Step-Studio/main/docs/mcp-skill.md) as `~/.claude/skills/ace-step-studio/SKILL.md`
- [CHANGELOG](https://github.com/timoncool/ACE-Step-Studio/blob/main/CHANGELOG.md)
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.