Skip to content

[DRAFT] feat: add VoiceStudio LXC installer - #2665

Closed
Estcool9 wants to merge 9 commits into
community-scripts:mainfrom
Estcool9:add-voicestudio-lxc
Closed

Estcool9 wants to merge 9 commits into
community-scripts:mainfrom
Estcool9:add-voicestudio-lxc

Conversation

@Estcool9

@Estcool9 Estcool9 commented Nov 15, 2026 •

Copy link
Copy Markdown

Description

Adds VoiceStudio as an unprivileged Debian 13 LXC, building the upstream web renderer and running its native FastAPI backend with systemd. It uses the upstream CPU-only setup path, generates a random administrator API key, and keeps user data and the Hugging Face model cache outside the application directory. The standard container updater rebuilds the source environment and web UI against the latest stable GitHub release. Installation details were checked against the upstream deploy/Dockerfile, docs/remote-gpu.md, electron/README.md, package.json, .python-version and pyproject.toml.

Related PR / Issue

New script submission; no existing issue found.

Prerequisites

  • Self-review completed against the current DevScripts AGENTS.md and script template.
  • Tested thoroughly. Runtime smoke tests and an update from the fork passed on an existing Proxmox CT. A fresh CT creation through the final helper flow and a real audio-generation test are still outstanding.
  • No breaking changes; this adds a new script set only.
  • No hardcoded credentials; a random key is generated and stored in a mode-600 EnvironmentFile.

Script

Name: VoiceStudio

Tool: Local-first AI voice studio for speech generation, voice cloning, dubbing, transcription and audiobooks. The LXC serves the browser UI and native inference backend on port 3900.

Category

  • CT — LXC container (ct/ + install/)
  • VM
  • Addon
  • PVE Tool
  • Other

Platform — tested on

  • Proxmox VE — runtime smoke testing has been done on an existing CT, but creation from this script has not yet been completed end to end.
  • Incus

Architecture

  • arm64 supported
  • arm64 not tested
  • arm64 not supported in this submission; the full native CPU dependency set has not been verified on arm64.

Type of change

  • New script

Code and Security Review

  • Uses the required CT, install and JSON metadata structure.
  • No hardcoded credentials.
  • No Docker; application is installed bare-metal.
  • No manual git clone or git pull; updates use the Community Scripts release helper.
  • Separate CODE-AUDIT.md and CONTRIBUTING.md guidance still needs maintainer confirmation. The current DevScripts AGENTS.md and README were used; the README confirms new scripts are developed and tested in DevScripts before moving to ProxmoxVE.

Additional Information

Static checks performed: bash -n on both shell scripts, JSON parsing, metadata/resource consistency, confirmation that the CT script defines only update_script(), and confirmation that persistent data is outside the clean-deploy application directory.

Runtime smoke tests were performed on Proxmox VE 9.2 using an existing unprivileged Debian 13 CT (6 vCPU, 12 GiB RAM, 50 GiB disk). VoiceStudio 0.5.8 started under systemd; /health returned status=ok with CPU selected; the web root and OpenAPI endpoints returned HTTP 200; the API-key-to-session exchange returned HTTP 201 and an authenticated admin endpoint returned HTTP 200. The Nginx Proxy Manager host could also reach the CT on port 3900.

The update function was exercised using the shared Core updater with the fork branch selected explicitly. It fetched release 0.5.8, rebuilt the web interface and Python environment, restarted the service and completed cleanup. The generated /bin/update on an existing CT still points to the upstream DevScripts main branch; that is expected before the PR is merged. The fork-branch override was used for the update test.

No fresh LXC has yet been created by running this CT script end to end, and no TTS model has been installed or audio-generation test performed. VoiceStudio currently reports OmniVoice as missing. Its bundled model record classifies the weights as non-commercial (CC-BY-NC, version unspecified), identifies separate Higgs Audio 2 and Meta Llama 3 terms, and marks component closure, upstream access and output terms as unresolved. The model-install API also reports that no reviewed production rows are currently enabled. No model licence acknowledgement was recorded and no model weights were downloaded. This is why the live-install and audio test boxes remain unchecked.

The initial submission is amd64-only. Debian 13 is the default. The installer uses Python 3.11, matching the upstream .python-version pin, and the upstream OMNIVOICE_TORCH_VARIANT=cpu setup path.

Application Requirements

  • Application age meets the six-month minimum (upstream repository creation checked as 2026-04-09).
  • Actively maintained, with recent commits and releases checked on 2026-11-15.
  • More than 1,000 GitHub stars at the time of review.
  • Official GitHub release tarballs are available and are used by the Community Scripts release helper.
  • I understand that acceptance remains at the discretion of the Community Scripts maintainers.

Sources

AI Assistance

AI was used to research the upstream installation, draft the scripts and check them against the current DevScripts instructions. The scripts and runtime checks were then reviewed against the live Proxmox test container.

GPT-6 (high reasoning) was used for research, implementation review, static validation and the live update/runtime smoke checks.

@github-actions

Copy link
Copy Markdown
Contributor

Try this script

ct/voicestudio.sh, run in the Proxmox VE shell or on an Incus host:

COMMUNITY_SCRIPTS_URL=https://raw.githubusercontent.com/Estcool9/DevScripts-voicestudio/add-voicestudio-lxc \
bash -c "$(curl -fsSL https://raw.githubusercontent.com/Estcool9/DevScripts-voicestudio/add-voicestudio-lxc/ct/voicestudio.sh)"

COMMUNITY_SCRIPTS_URL is not optional for ct/. Fetching the ct/ script from a
branch does not tell the engine where that branch is — with bash -c "$(curl …)"
there is no file on disk for the scripts root to be derived from, so it would fall
back to upstream main and look for the install script there.

Against a core branch as well

Add COMMUNITY_SCRIPTS_CORE_URL=https://raw.githubusercontent.com/OWNER/core/BRANCH
to test an engine change at the same time. The two resolve independently.

Useful flags while testing

dev_mode=net logs every fetch with status and duration, so you can confirm the
branch is really being used. dev_mode=keep stops a failed build from deleting
the container along with the evidence.

@github-actions

Copy link
Copy Markdown
Contributor

👋 @Estcool9 this PR adds new file(s) under ct/ or install/, but the 🆕 New script box in the Type of Change section isn't checked:

  • ct/voicestudio.sh
  • install/voicestudio-install.sh

If this is a new script submission, please check that box and fill out the 📦 Application Requirements section — it's required for review.
If this isn't actually a new script (e.g. a rename or a shared helper file), you can ignore this message.

@github-actions

Copy link
Copy Markdown
Contributor

⚠️ Not ready for review yet

@Estcool9, this pull request is missing the following:

  • The description does not use the DevScripts PR template (the 📦 Application Requirements section is missing). Paste .github/pull_request_template.md into the description and fill it in; your own write-up can go under Description.

Edit the PR description to fix it — no new PR needed. This check runs again on every edit and clears itself once everything is in place.

Until then the PR is marked stale and will be closed in 7 days.


@community-scripts/contributor please take a look: close it, comment, or set keep-open to keep it.

@github-actions github-actions Bot added requirements missing PR description misses template items; closes after 7 days unless fixed or keep-open stale labels Nov 15, 2026
@michelroegl-brunner

Copy link
Copy Markdown
Member

Use the correct template, then repoen the pr.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

new script requirements missing PR description misses template items; closes after 7 days unless fixed or keep-open stale

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants