Palash Debnathandmergetest
4e5d795832
Two recon findings from the reset work, fixed properly (whole class + tests +
docs), plus the destructive reset path is now exercised end-to-end.
1. ~/.config/omnivoice/env survived every uninstall. The app persists the
model-cache location (and a possible HF_TOKEN) there via
backend/core/user_env.py, but the in-app "Remove all data" (uninstall.rs),
uninstall.sh, and uninstall.ps1 all walked past it — so a reinstall silently
inherited the old file and redirected downloads to a maybe-deleted location.
All three now remove it. It's the same expanduser("~/.config/omnivoice/env")
path on every OS, so the Windows script uses %USERPROFILE%\.config\omnivoice.
is_recognizably_ours accepts it (contains "omnivoice"); docs tables updated.
2. Disk usage measured the wrong engines dir. storage_report.default_engines_dir()
returned backend/engines (built-in engine *modules*, no venvs), while sidecar
installs live in DATA_DIR/engines/<id>. So a multi-GB IndexTTS-2 install was
invisible in the engine-venv category and rolled into data/"other". Now points
at DATA_DIR/engines and sizes the WHOLE install (venv + checkout + weights),
with the data category claiming that subtree so it isn't double-counted.
Reset hardening: extracted purge_scopes() as a pure fs function (no AppHandle),
so the actual delete loop runs in tests against a real on-disk install tree —
"everything" wipes the install but spares the venv/foreign temp/sibling folders,
a settings reset keeps content+config+models, and a poisoned data_dir="$HOME"
deletes NOTHING. This is the live drive-through of the destructive path, minus
the GUI.
Also: gitignore the node_modules symlink form (the directory rule node_modules/
never matched a worktree symlink, so it kept slipping into commits).
Tests: Rust 78 (6 new), storage_report 20 (2 new incl. once-not-twice count +
default-dir guard), frontend 1207, i18n probe green, format+lint clean.
Co-authored-by: mergetest <nizam4103@gmail.com>
8.8 KiB
Uninstalling OmniVoice Studio
OmniVoice is fully local — it has no accounts, no cloud state, and no background services. Removing it is just deleting the app plus the folders it wrote on your machine. This page lists every one of those folders per platform, and ships a script that finds and removes them for you (with a dry-run first).
TL;DR (the space hogs): the two folders worth deleting are the model cache (the Hugging Face weights — several GB) and the managed Python environment (
project/.venv— a few GB). Everything else is small.
In the app (easiest — no repo needed)
Settings → Storage → Remove all data. It lists every folder this install
owns with its real size, lets you opt in (separately) to the shared Hugging Face
model cache, asks you to type DELETE, then removes everything and quits.
This is the right path if you installed the .dmg / .msi / AppImage — you don't have the repo, so the script below isn't available to you.
You may not need to uninstall. Right above it, Reset & remove does the same job at any scale you like — and leaves you with a working app instead of no app. See Resetting below.
Resetting instead of uninstalling
Settings → Storage → Reset & remove puts part — or all — of OmniVoice back to how it shipped, without removing the app. Every option shows its real size before you commit, and the app restarts itself when it's done.
| Option | What it removes | What it keeps |
|---|---|---|
| UI preferences only | Theme, layout, language, dub settings | Everything on disk |
| All settings | The above, plus saved settings on disk (engine choices, voice defaults) | Voices, projects, audio, models |
| Downloaded assets & models | Model weights, sidecar engines, audio tools, caches | Everything you made |
| Everything OmniVoice did | All of the above, plus voices, projects, generated audio, history, logs | The app itself, and the Python environment it runs on |
"Choose exactly what to remove" opens the same list as individual checkboxes, so you can drop just the model weights, just a wedged sidecar engine, or just the history — whatever is actually wrong.
Two things it deliberately does not touch:
- Your storage locations. If you pointed OmniVoice at a custom data or model directory, a settings reset keeps that pointer. Clearing it would strand gigabytes of already-downloaded weights at a path the app no longer looks in.
- The managed Python environment. "Everything OmniVoice did" still leaves you with a working app that restarts on the first-run screen. If you want the interpreter gone too, that's Remove all data — the section above.
The shared Hugging Face model cache is called out separately wherever it applies: on macOS and Linux it's the standard cache other AI tools use too, so removing it may delete models OmniVoice never downloaded. (On Windows, and in portable installs, the cache is OmniVoice's own — there's nothing to share, and the app says so.)
The one-command uninstaller (from a clone)
From a clone or the source tarball:
# macOS / Linux — prints what it WOULD delete, with sizes, and stops:
scripts/uninstall.sh
# actually delete, after you've read the dry-run:
scripts/uninstall.sh --yes
# Windows (PowerShell) — dry-run, then delete:
powershell -ExecutionPolicy Bypass -File scripts\uninstall.ps1
powershell -ExecutionPolicy Bypass -File scripts\uninstall.ps1 -Yes
The script honors your custom locations: if you set OMNIVOICE_DATA_DIR,
OMNIVOICE_CACHE_DIR, HF_HOME, or HF_HUB_CACHE (or picked custom
data/model folders during setup), export the same variables before running it
and it will target those instead of the defaults. It never touches anything
outside the OmniVoice folders, and it does not delete the app binary itself
(see "Remove the app" below) — so it's safe to run even if you only want to
reclaim disk space and keep the app installed.
What OmniVoice writes, and where
Four kinds of data, in up to four locations:
| What | Size | Notes |
|---|---|---|
| Model cache (Hugging Face weights) | GBs | The big one. Shared HF cache — see the caveat below. |
Managed Python env (project/.venv) |
GBs | Rebuilt automatically if you reinstall. |
| App data (voices, projects, DB, generated audio, logs) | small–MBs | Your voice profiles and history live here. |
App config + logs (config.json, window state, Tauri logs) |
tiny | Settings and desktop-shell logs. |
macOS
~/Library/Application Support/OmniVoice/ ← app data (voices, projects, omnivoice.db, outputs, omnivoice.log)
~/Library/Application Support/com.debpalash.omnivoice-studio/ ← config.json + the managed Python env (project/.venv)
~/Library/Logs/OmniVoice/ ← backend logs (backend.log, backend_err.log)
~/Library/Logs/com.debpalash.omnivoice-studio/ ← desktop-shell log (tauri.log)
~/.config/omnivoice/ ← saved env file (cache location, HF token)
~/.cache/huggingface/ ← model weights (shared HF cache — see caveat)
Linux (AppImage / .deb)
~/.omnivoice/ ← app data (voices, projects, omnivoice.db, outputs, omnivoice.log)
~/.local/share/com.debpalash.omnivoice-studio/ ← config.json, shell logs, AND the managed Python env (project/.venv)
~/.local/state/OmniVoice/ ← backend logs (backend.log, backend_err.log)
~/.config/omnivoice/ ← saved env file (cache location, HF token)
~/.cache/huggingface/ ← model weights (shared HF cache — see caveat)
Windows
%APPDATA%\OmniVoice\ ← app data (voices, projects, omnivoice.db, outputs, omnivoice.log)
%LOCALAPPDATA%\com.debpalash.omnivoice-studio\ ← config.json, shell logs, AND the managed Python env (project\.venv)
%LOCALAPPDATA%\OmniVoice\Logs\ ← backend logs (backend.log, backend_err.log)
%USERPROFILE%\.config\omnivoice\ ← saved env file (cache location, HF token)
%LOCALAPPDATA%\OmniVoice\hf_cache\ ← model weights (OmniVoice uses a short path here to dodge MAX_PATH)
On Windows, if HF_HOME isn't set, OmniVoice redirects the model cache to
%LOCALAPPDATA%\OmniVoice\hf_cache (instead of ~/.cache/huggingface) so deep
model paths don't hit the 260-character MAX_PATH limit.
Custom / portable locations
- Custom folders: if you chose a custom data or model folder in setup (or
set
OMNIVOICE_DATA_DIR/OMNIVOICE_CACHE_DIR/HF_HOME/HF_HUB_CACHE), your data is there instead of the defaults above. - Portable mode: everything lives in an
OmniVoiceStudio-Data/folder next to the app binary — delete that one folder and you're done. - Optional engine sidecars: if you installed IndexTTS-2, CosyVoice, or another sidecar engine, its isolated venv lives under the app-config folder above and is removed with it.
Shared HF cache caveat:
~/.cache/huggingface/is the standard Hugging Face cache, shared by any tool that useshuggingface_hub(other ML apps,transformers, etc.). If you use other Hugging Face tools, deleting the whole folder removes their cached models too. To remove only OmniVoice's models, delete themodels--*subfolders you recognize under~/.cache/huggingface/hub/, or just let it be — it's only a cache and any tool re-downloads what it needs. The uninstaller script prints the cache size and asks about it separately for this reason.
Remove the app itself
The steps above clear the data; removing the installed app is the normal per-platform step:
- macOS: drag OmniVoice Studio.app from
/Applicationsto the Trash. - Windows: Settings → Apps → Installed apps → OmniVoice Studio → Uninstall (or via "Add or remove programs").
- Linux (AppImage): delete the
.AppImagefile. If you integrated it into your menu (e.g. with AppImageLauncher or a hand-written.desktopfile), also remove~/.local/share/applications/*omnivoice*.desktopand any icon under~/.local/share/icons/. - Linux (.deb):
sudo apt remove omnivoice-studio(this removes the program; your data folders above are user data and are left in place — delete them with the script or by hand).
Reinstalling later
Nothing above is required before reinstalling — a fresh install rebuilds the
Python env and re-downloads models on demand. Keep ~/Library/Application Support/OmniVoice/ (macOS) / ~/.omnivoice/ (Linux) / %APPDATA%\OmniVoice\
(Windows) if you want to preserve your voice profiles and projects across a
reinstall; delete it too for a truly clean slate.