Add Voicebox app definition #13

Merged

backmeupplz merged 1 commit from myg-kaneo-10-voicebox into master

2026-05-25 16:23:18 +00:00

backmeupplz commented

2026-05-25 16:06:03 +00:00

(Migrated from github.com)

Summary

Add Voicebox to the MyGround app catalog with metadata, SVG icon, and public README/website catalog entries.
Add a source-build Docker Compose template based on the official Voicebox Docker deployment: port 17493, persistent generated audio, app data, and Hugging Face model cache storage.
Add configurable log level, model cache path, CORS origins, CPU/memory limits, GPU toggle support, bind-mount permission init, and post-install security/resource/update notes.
Add registry tests covering Voicebox metadata, storage, GPU support, compose details, and icon style.

Validation

make build-web - passed; note: local Bun 1.1.42 could not parse the current lockfile version, ignored it, installed dependencies, built web/dist, and I restored the generated web/bun.lock change.
rustfmt --check src/registry.rs - passed.
docker info --format ... - passed, Docker Server 27.5.1 available.
docker compose -f apps/voicebox.yml config with representative MyGround env/storage variables - passed; verified expanded port, healthcheck, bind mounts, resource limits, and one-shot init service.
cargo test voicebox - passed.
cargo test - passed: 285 lib tests, 6 main tests, 83 API tests; 12 backup e2e tests ignored by default.
cd web && bunx vitest run - passed: 17 files, 115 tests.
git diff --check - passed.

Manual QA

Manual Docker install smoke was not run because Voicebox has no prebuilt image yet and the official path requires building the multi-stage source image plus heavy Python/TTS dependencies. Reviewer/OpenClaw smoke path: install Voicebox from this branch, keep the default localhost/Tailscale-only exposure, wait for the source build and first model download, open the configured MyGround URL on port 17493, confirm /health succeeds, create or load a voice profile, generate a short test clip, restart the app, and verify generated audio plus model cache persist without re-download.

Sources checked: https://docs.voicebox.sh/overview/docker and https://github.com/jamiepine/voicebox.

## Summary - Add Voicebox to the MyGround app catalog with metadata, SVG icon, and public README/website catalog entries. - Add a source-build Docker Compose template based on the official Voicebox Docker deployment: port 17493, persistent generated audio, app data, and Hugging Face model cache storage. - Add configurable log level, model cache path, CORS origins, CPU/memory limits, GPU toggle support, bind-mount permission init, and post-install security/resource/update notes. - Add registry tests covering Voicebox metadata, storage, GPU support, compose details, and icon style. ## Validation - `make build-web` - passed; note: local Bun 1.1.42 could not parse the current lockfile version, ignored it, installed dependencies, built `web/dist`, and I restored the generated `web/bun.lock` change. - `rustfmt --check src/registry.rs` - passed. - `docker info --format ...` - passed, Docker Server 27.5.1 available. - `docker compose -f apps/voicebox.yml config` with representative MyGround env/storage variables - passed; verified expanded port, healthcheck, bind mounts, resource limits, and one-shot init service. - `cargo test voicebox` - passed. - `cargo test` - passed: 285 lib tests, 6 main tests, 83 API tests; 12 backup e2e tests ignored by default. - `cd web && bunx vitest run` - passed: 17 files, 115 tests. - `git diff --check` - passed. ## Manual QA Manual Docker install smoke was not run because Voicebox has no prebuilt image yet and the official path requires building the multi-stage source image plus heavy Python/TTS dependencies. Reviewer/OpenClaw smoke path: install Voicebox from this branch, keep the default localhost/Tailscale-only exposure, wait for the source build and first model download, open the configured MyGround URL on port 17493, confirm `/health` succeeds, create or load a voice profile, generate a short test clip, restart the app, and verify generated audio plus model cache persist without re-download. Sources checked: https://docs.voicebox.sh/overview/docker and https://github.com/jamiepine/voicebox.

backmeupplz commented

2026-05-25 16:23:16 +00:00

(Migrated from github.com)

Review/QA for MYG-10 ("Add Voicebox as a MyGround app"):

Checked the app definition against upstream Voicebox Docker compose/docs and MyGround registry patterns. No blocking findings.

Validation run:

docker compose -f apps/voicebox.yml config with representative MyGround env/storage variables
cargo test voicebox -- --nocapture
cargo run -- app list | rg -i 'voicebox|available|ai voice|transcription'
git diff --check
cargo test
cd web && bunx vitest run
make build-web (passed; local Bun 1.1.42 still rewrites/ignores the current lockfile format, restored the generated web/bun.lock change)

GitHub will not let me approve the PR because it is authored by the same account, but it passes review and QA.

Review/QA for MYG-10 ("Add Voicebox as a MyGround app"): Checked the app definition against upstream Voicebox Docker compose/docs and MyGround registry patterns. No blocking findings. Validation run: - `docker compose -f apps/voicebox.yml config` with representative MyGround env/storage variables - `cargo test voicebox -- --nocapture` - `cargo run -- app list | rg -i 'voicebox|available|ai voice|transcription'` - `git diff --check` - `cargo test` - `cd web && bunx vitest run` - `make build-web` (passed; local Bun 1.1.42 still rewrites/ignores the current lockfile format, restored the generated `web/bun.lock` change) GitHub will not let me approve the PR because it is authored by the same account, but it passes review and QA.