PyPI - studio-console - Versions diffs - 1.3.4__tar.gz → 1.3.5__tar.gz - Mend

studio-console 1.3.4tar.gz → 1.3.5tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (29) hide show

{studio_console-1.3.4 → studio_console-1.3.5}/PKG-INFO RENAMED Viewed

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: studio-console
-Version: 1.3.4
+Version: 1.3.5
 Summary: Studio operator management console
 Project-URL: Homepage, https://github.com/selfhosthub/studio-console
 Project-URL: Issues, https://github.com/selfhosthub/studio-console/issues
@@ -24,9 +24,9 @@ Operator CLI for managing a self-hosted [Studio](https://github.com/selfhosthub/
 Studio ships three images. Most operators want **Split**.
-- **Split** — multi-container compose stack. Console runs on the host and manages `docker compose`. **This README documents Split.**
-- **Core** (`studio-core`) — single container bundling API + UI; you bring an external Postgres. Console runs inside the container.
-- **Full** (`studio-full` / `studio-standalone`) — single container with bundled Postgres.
+- **Split**: multi-container compose stack. Console runs on the host and manages `docker compose`. **This README documents Split.**
+- **Core** (`studio-core`): single container bundling API + UI; you bring an external Postgres. Console runs inside the container.
+- **Full** (`studio-full`): single container with bundled Postgres.
 In Core and Full the container entrypoint provisions Studio on first boot; the console is a diagnostic + ops tool.
@@ -36,6 +36,7 @@ In Core and Full the container entrypoint provisions Studio on first boot; the c
 - [Quick Start](#quick-start)
 - [Run the Full image](#run-the-full-image)
+- [Run the Core image](#run-the-core-image)
 - [Daily Operations](#daily-operations)
 - [Upgrading](#upgrading)
 - [CLI Reference](#cli-reference)
@@ -52,15 +53,17 @@ In Core and Full the container entrypoint provisions Studio on first boot; the c
 ```sh
 # 1. Install uv (one-time)
 curl -LsSf https://astral.sh/uv/install.sh | sh
-source ~/.bashrc           # macOS default shell is zsh — use ~/.zshrc instead
+source ~/.bashrc           # macOS default shell is zsh, use ~/.zshrc instead
-# 2. Install studio-console
-uv tool install https://github.com/selfhosthub/studio-console/releases/download/v1.3.4/studio_console-1.3.4-py3-none-any.whl
+# 2. Install studio-console (isolated CLI, recommended)
+uv tool install studio-console
 # 3. Run
 studio-console
 ```
+No uv and don't want it? `pip install studio-console` works anywhere Python 3.8+ is available. It installs into the active environment instead of an isolated one, so a virtualenv is a good idea.
 On first run with no `~/.studio/.env`, the wizard launches. Walk through the sections, save, then go to **Services → Start all**.
 The wizard creates `~/.studio/` containing `.env` (0600), `docker-compose.yml`, `nginx/studio.conf`, and four data subdirs:
@@ -69,15 +72,15 @@ The wizard creates `~/.studio/` containing `.env` (0600), `docker-compose.yml`,
 ~/.studio/
 ├── .env, .bootstrapped, docker-compose.yml, nginx/
 ├── db/         postgres data         (SHS_DB_DATA)
-├── storage/    orgs, uploads, outputs (SHS_STORAGE_ROOT — mounted at /workspace in containers)
+├── storage/    orgs, uploads, outputs (SHS_STORAGE_ROOT, mounted at /workspace in containers)
 ├── models/     model files            (SHS_MODELS_ROOT)
 └── backups/    local DB dumps         (SHS_BACKUP_ROOT)
 ```
-Each data subdir is its own env var, so a cloud deploy can repoint individual roots at CloudSQL, GCS, or a network volume without changing code. **Protect this directory and never delete `.env`** — two of its values cannot be regenerated:
+Each data subdir is its own env var, so a cloud deploy can repoint individual roots at CloudSQL, GCS, or a network volume without changing code. **Protect this directory and never delete `.env`.** Two of its values cannot be regenerated:
-- `SHS_CREDENTIAL_ENCRYPTION_KEY` — losing it makes all stored provider API keys unrecoverable.
-- `POSTGRES_PASSWORD` — set on first DB init; the live database keeps using this value, so a regenerated one will lock the API out.
+- `SHS_CREDENTIAL_ENCRYPTION_KEY`: losing it makes all stored provider API keys unrecoverable.
+- `POSTGRES_PASSWORD`: set on first DB init; the live database keeps using this value, so a regenerated one will lock the API out.
 Back both up separately (password manager, secrets vault) so you can recover if the host or `.env` is ever lost.
@@ -96,23 +99,23 @@ Full list of supported environment variables in [docs/env-vars.md](docs/env-vars
 ## Run the Full image
-The **Full** image (`studio-full`) is a single self-contained container — bundled Postgres, API, UI, and workers under supervisord. No Compose stack, no external database. `launch-full` runs it on your machine and drops you into its in-container console.
+The **Full** image (`studio-full`) is a single self-contained container with bundled Postgres, API, UI, and workers under supervisord. No Compose stack, no external database. `launch-full` runs it on your machine and drops you into its in-container console.
 ```sh
-# Launch (workspace defaults to ~/.studio, shared with Split)
-studio-console launch-full --tag 1.0.0
+# Launch (defaults to the latest tag; workspace defaults to ~/.studio, shared with Split)
+studio-console launch-full
-# Use a separate workspace to run Full alongside a Split install
-studio-console launch-full --tag 1.0.0 --workspace ~/.studio-full
+# Pin a tag and/or use a separate workspace to run Full alongside a Split install
+studio-console launch-full --tag 1.2.4 --workspace ~/.studio-full
 ```
-The workspace (mounted at `/workspace`) holds the generated `.env`, the Postgres data dir, and org files — it persists across restarts. You're prompted once for a supervisor username/password; the console remembers them and re-injects on every launch.
+The workspace (mounted at `/workspace`) holds the generated `.env`, the Postgres data dir, and org files; it persists across restarts. You're prompted once for a supervisor username/password; the console remembers them and re-injects on every launch.
 > **Do not point Full at an existing Split `~/.studio` workspace** unless you mean to share it. Split and Full share the same `.env`/encryption key but **must never run concurrently** against the same data. Use `--workspace` to keep them separate.
 ### Re-entering the console after you exit
-Exiting the console **does not stop the container** — it keeps running. `launch-full` won't re-attach to an already-running container; instead, exec back in from the host:
+Exiting the console **does not stop the container**; it keeps running. `launch-full` won't re-attach to an already-running container; instead, exec back in from the host:
 ```sh
 docker exec -it studio-full studio-console        # re-open the in-container menu
@@ -127,6 +130,38 @@ Inside the container the console manages services via `supervisorctl` (health, r
 ---
+## Run the Core image
+The **Core** image (`studio-core`) is the same single-container API + UI as Full, but without a bundled database. You bring your own Postgres (a managed cloud instance, a RunPod sidecar, any reachable server). `launch-core` runs it on your machine and drops you into its in-container console.
+Point Core at your database first, then launch:
+```sh
+# Set the external database URL (prompts if you omit it)
+studio-console core-db-url postgresql+asyncpg://user:pass@host:5432/studio
+# Launch (defaults to the latest tag; workspace defaults to ~/.studio-core)
+studio-console launch-core
+# Pin a tag and/or use a separate workspace
+studio-console launch-core --tag 1.2.4 --workspace ~/.studio-core
+```
+Precedence for the database URL: a saved value from `core-db-url`, then the `SHS_DATABASE_URL` environment variable, then an interactive prompt. Export the env var to boot Core unattended against a remote Postgres with no prompt:
+```sh
+SHS_DATABASE_URL=postgresql+asyncpg://user:pass@host:5432/studio studio-console launch-core
+```
+Re-enter and run one-off commands the same way as Full, swapping the container name:
+```sh
+docker exec -it studio-core studio-console        # re-open the in-container menu
+docker exec studio-core studio-console health
+```
+---
 ## Daily Operations
 ```
@@ -153,33 +188,15 @@ Backups land in `~/.studio/backups/studio-YYYYMMDD_HHMMSS/`.
 Images → Upgrade
 ```
-Pulls the latest tag from GHCR, updates `SHS_STUDIO_VERSION` in `.env`, and restarts all services. If the new version's major doesn't match your database's major, the upgrade is blocked with a clear message — see [docs/architecture.md](docs/architecture.md#major-version-boundary-detection).
+Pulls the latest tag from GHCR, updates `SHS_STUDIO_VERSION` in `.env`, and restarts all services. If the new version's major doesn't match your database's major, the upgrade is blocked with a clear message. See [docs/architecture.md](docs/architecture.md#major-version-boundary-detection).
 **studio-console itself:**
 ```bash
-uv tool install --force <new-release-url>
-```
-Latest release is always at [github.com/selfhosthub/studio-console/releases/latest](https://github.com/selfhosthub/studio-console/releases/latest).
----
-## Cutting a release
-**One command. Never tag, bump `VERSION`, or run `gh release` by hand** — the script does all of it (bumps `VERSION`, updates the README install URL, commits, pushes, tags, builds the wheel, publishes the release with the correct title).
-```bash
-# 1. Land your fix on main first (commit + push as normal).
-# 2. Then cut the release — pick the bump size:
-scripts/release-console.sh patch    # bug fix      1.0.0 → 1.0.1
-scripts/release-console.sh minor    # new feature  1.0.0 → 1.1.0
-scripts/release-console.sh major    # breaking     1.0.0 → 2.0.0
+studio-console self-update
 ```
-Preview without touching anything: add `--dry-run`. Override the notes with `--message "..."`. Re-cut a release you already published (e.g. it shipped before a fix landed): `scripts/release-console.sh <X.Y.Z> --force`.
-After release, bump the console pin in the Studio versions file to the new version so the next core/full image build bakes it in.
+`self-update` detects how Console was installed (uv, pip, or Homebrew) and upgrades in place.
 ---
@@ -200,6 +217,7 @@ studio-console restore [path]           # restore from backup directory
 studio-console links                    # print service URLs
 studio-console config                   # show current .env values
 studio-console config set KEY VALUE     # set a single .env value
+studio-console config unset KEY         # remove a single .env value
 studio-console workers                  # list/scale workers
 studio-console reset-password           # reset super admin password
 studio-console wizard                   # re-run setup wizard
@@ -208,18 +226,20 @@ studio-console self-update              # upgrade studio-console itself
 studio-console version                  # print version
 studio-console launch-full [--tag T] [--workspace DIR]   # run the Full single-container image on the host
+studio-console launch-core [--tag T] [--workspace DIR]   # run the Core single-container image (external Postgres)
+studio-console core-db-url [URL]        # set Core's external database URL for the next launch-core
 ```
-For the Full image, run operational subcommands through `docker exec studio-full studio-console <cmd>` — see [Run the Full image](#run-the-full-image).
+For the Full or Core image, run operational subcommands through `docker exec studio-full studio-console <cmd>` (or `studio-core`). See [Run the Full image](#run-the-full-image) and [Run the Core image](#run-the-core-image).
 ---
 ## Further reading
-- **[Architecture](docs/architecture.md)** — file layout, state flow, Compose wiring, first-boot, internals (wizard vs init, orphan workers, backup format, major-version detection, deployment contexts).
-- **[Environment variables](docs/env-vars.md)** — every `SHS_*` and supporting var.
-- **[Public hostname topology](docs/topology.md)** — single vs split hostnames, Cloudflare tunnel + Access, IP restrictions.
-- **[VPS + RunPod deployments](docs/vps-runpod.md)** — hybrid setups with GPU worker pods on RunPod.
+- **[Architecture](docs/architecture.md)**: file layout, state flow, Compose wiring, first-boot, internals (wizard vs init, orphan workers, backup format, major-version detection, deployment contexts).
+- **[Environment variables](docs/env-vars.md)**: every `SHS_*` and supporting var.
+- **[Public hostname topology](docs/topology.md)**: single vs split hostnames, Cloudflare tunnel + Access, IP restrictions.
+- **[VPS + RunPod deployments](docs/vps-runpod.md)**: hybrid setups with GPU worker pods on RunPod.
 ---

{studio_console-1.3.4 → studio_console-1.3.5}/README.md RENAMED Viewed

@@ -6,9 +6,9 @@ Operator CLI for managing a self-hosted [Studio](https://github.com/selfhosthub/
 Studio ships three images. Most operators want **Split**.
-- **Split** — multi-container compose stack. Console runs on the host and manages `docker compose`. **This README documents Split.**
-- **Core** (`studio-core`) — single container bundling API + UI; you bring an external Postgres. Console runs inside the container.
-- **Full** (`studio-full` / `studio-standalone`) — single container with bundled Postgres.
+- **Split**: multi-container compose stack. Console runs on the host and manages `docker compose`. **This README documents Split.**
+- **Core** (`studio-core`): single container bundling API + UI; you bring an external Postgres. Console runs inside the container.
+- **Full** (`studio-full`): single container with bundled Postgres.
 In Core and Full the container entrypoint provisions Studio on first boot; the console is a diagnostic + ops tool.
@@ -18,6 +18,7 @@ In Core and Full the container entrypoint provisions Studio on first boot; the c
 - [Quick Start](#quick-start)
 - [Run the Full image](#run-the-full-image)
+- [Run the Core image](#run-the-core-image)
 - [Daily Operations](#daily-operations)
 - [Upgrading](#upgrading)
 - [CLI Reference](#cli-reference)
@@ -34,15 +35,17 @@ In Core and Full the container entrypoint provisions Studio on first boot; the c
 ```sh
 # 1. Install uv (one-time)
 curl -LsSf https://astral.sh/uv/install.sh | sh
-source ~/.bashrc           # macOS default shell is zsh — use ~/.zshrc instead
+source ~/.bashrc           # macOS default shell is zsh, use ~/.zshrc instead
-# 2. Install studio-console
-uv tool install https://github.com/selfhosthub/studio-console/releases/download/v1.3.4/studio_console-1.3.4-py3-none-any.whl
+# 2. Install studio-console (isolated CLI, recommended)
+uv tool install studio-console
 # 3. Run
 studio-console
 ```
+No uv and don't want it? `pip install studio-console` works anywhere Python 3.8+ is available. It installs into the active environment instead of an isolated one, so a virtualenv is a good idea.
 On first run with no `~/.studio/.env`, the wizard launches. Walk through the sections, save, then go to **Services → Start all**.
 The wizard creates `~/.studio/` containing `.env` (0600), `docker-compose.yml`, `nginx/studio.conf`, and four data subdirs:
@@ -51,15 +54,15 @@ The wizard creates `~/.studio/` containing `.env` (0600), `docker-compose.yml`,
 ~/.studio/
 ├── .env, .bootstrapped, docker-compose.yml, nginx/
 ├── db/         postgres data         (SHS_DB_DATA)
-├── storage/    orgs, uploads, outputs (SHS_STORAGE_ROOT — mounted at /workspace in containers)
+├── storage/    orgs, uploads, outputs (SHS_STORAGE_ROOT, mounted at /workspace in containers)
 ├── models/     model files            (SHS_MODELS_ROOT)
 └── backups/    local DB dumps         (SHS_BACKUP_ROOT)
 ```
-Each data subdir is its own env var, so a cloud deploy can repoint individual roots at CloudSQL, GCS, or a network volume without changing code. **Protect this directory and never delete `.env`** — two of its values cannot be regenerated:
+Each data subdir is its own env var, so a cloud deploy can repoint individual roots at CloudSQL, GCS, or a network volume without changing code. **Protect this directory and never delete `.env`.** Two of its values cannot be regenerated:
-- `SHS_CREDENTIAL_ENCRYPTION_KEY` — losing it makes all stored provider API keys unrecoverable.
-- `POSTGRES_PASSWORD` — set on first DB init; the live database keeps using this value, so a regenerated one will lock the API out.
+- `SHS_CREDENTIAL_ENCRYPTION_KEY`: losing it makes all stored provider API keys unrecoverable.
+- `POSTGRES_PASSWORD`: set on first DB init; the live database keeps using this value, so a regenerated one will lock the API out.
 Back both up separately (password manager, secrets vault) so you can recover if the host or `.env` is ever lost.
@@ -78,23 +81,23 @@ Full list of supported environment variables in [docs/env-vars.md](docs/env-vars
 ## Run the Full image
-The **Full** image (`studio-full`) is a single self-contained container — bundled Postgres, API, UI, and workers under supervisord. No Compose stack, no external database. `launch-full` runs it on your machine and drops you into its in-container console.
+The **Full** image (`studio-full`) is a single self-contained container with bundled Postgres, API, UI, and workers under supervisord. No Compose stack, no external database. `launch-full` runs it on your machine and drops you into its in-container console.
 ```sh
-# Launch (workspace defaults to ~/.studio, shared with Split)
-studio-console launch-full --tag 1.0.0
+# Launch (defaults to the latest tag; workspace defaults to ~/.studio, shared with Split)
+studio-console launch-full
-# Use a separate workspace to run Full alongside a Split install
-studio-console launch-full --tag 1.0.0 --workspace ~/.studio-full
+# Pin a tag and/or use a separate workspace to run Full alongside a Split install
+studio-console launch-full --tag 1.2.4 --workspace ~/.studio-full
 ```
-The workspace (mounted at `/workspace`) holds the generated `.env`, the Postgres data dir, and org files — it persists across restarts. You're prompted once for a supervisor username/password; the console remembers them and re-injects on every launch.
+The workspace (mounted at `/workspace`) holds the generated `.env`, the Postgres data dir, and org files; it persists across restarts. You're prompted once for a supervisor username/password; the console remembers them and re-injects on every launch.
 > **Do not point Full at an existing Split `~/.studio` workspace** unless you mean to share it. Split and Full share the same `.env`/encryption key but **must never run concurrently** against the same data. Use `--workspace` to keep them separate.
 ### Re-entering the console after you exit
-Exiting the console **does not stop the container** — it keeps running. `launch-full` won't re-attach to an already-running container; instead, exec back in from the host:
+Exiting the console **does not stop the container**; it keeps running. `launch-full` won't re-attach to an already-running container; instead, exec back in from the host:
 ```sh
 docker exec -it studio-full studio-console        # re-open the in-container menu
@@ -109,6 +112,38 @@ Inside the container the console manages services via `supervisorctl` (health, r
 ---
+## Run the Core image
+The **Core** image (`studio-core`) is the same single-container API + UI as Full, but without a bundled database. You bring your own Postgres (a managed cloud instance, a RunPod sidecar, any reachable server). `launch-core` runs it on your machine and drops you into its in-container console.
+Point Core at your database first, then launch:
+```sh
+# Set the external database URL (prompts if you omit it)
+studio-console core-db-url postgresql+asyncpg://user:pass@host:5432/studio
+# Launch (defaults to the latest tag; workspace defaults to ~/.studio-core)
+studio-console launch-core
+# Pin a tag and/or use a separate workspace
+studio-console launch-core --tag 1.2.4 --workspace ~/.studio-core
+```
+Precedence for the database URL: a saved value from `core-db-url`, then the `SHS_DATABASE_URL` environment variable, then an interactive prompt. Export the env var to boot Core unattended against a remote Postgres with no prompt:
+```sh
+SHS_DATABASE_URL=postgresql+asyncpg://user:pass@host:5432/studio studio-console launch-core
+```
+Re-enter and run one-off commands the same way as Full, swapping the container name:
+```sh
+docker exec -it studio-core studio-console        # re-open the in-container menu
+docker exec studio-core studio-console health
+```
+---
 ## Daily Operations
 ```
@@ -135,33 +170,15 @@ Backups land in `~/.studio/backups/studio-YYYYMMDD_HHMMSS/`.
 Images → Upgrade
 ```
-Pulls the latest tag from GHCR, updates `SHS_STUDIO_VERSION` in `.env`, and restarts all services. If the new version's major doesn't match your database's major, the upgrade is blocked with a clear message — see [docs/architecture.md](docs/architecture.md#major-version-boundary-detection).
+Pulls the latest tag from GHCR, updates `SHS_STUDIO_VERSION` in `.env`, and restarts all services. If the new version's major doesn't match your database's major, the upgrade is blocked with a clear message. See [docs/architecture.md](docs/architecture.md#major-version-boundary-detection).
 **studio-console itself:**
 ```bash
-uv tool install --force <new-release-url>
-```
-Latest release is always at [github.com/selfhosthub/studio-console/releases/latest](https://github.com/selfhosthub/studio-console/releases/latest).
----
-## Cutting a release
-**One command. Never tag, bump `VERSION`, or run `gh release` by hand** — the script does all of it (bumps `VERSION`, updates the README install URL, commits, pushes, tags, builds the wheel, publishes the release with the correct title).
-```bash
-# 1. Land your fix on main first (commit + push as normal).
-# 2. Then cut the release — pick the bump size:
-scripts/release-console.sh patch    # bug fix      1.0.0 → 1.0.1
-scripts/release-console.sh minor    # new feature  1.0.0 → 1.1.0
-scripts/release-console.sh major    # breaking     1.0.0 → 2.0.0
+studio-console self-update
 ```
-Preview without touching anything: add `--dry-run`. Override the notes with `--message "..."`. Re-cut a release you already published (e.g. it shipped before a fix landed): `scripts/release-console.sh <X.Y.Z> --force`.
-After release, bump the console pin in the Studio versions file to the new version so the next core/full image build bakes it in.
+`self-update` detects how Console was installed (uv, pip, or Homebrew) and upgrades in place.
 ---
@@ -182,6 +199,7 @@ studio-console restore [path]           # restore from backup directory
 studio-console links                    # print service URLs
 studio-console config                   # show current .env values
 studio-console config set KEY VALUE     # set a single .env value
+studio-console config unset KEY         # remove a single .env value
 studio-console workers                  # list/scale workers
 studio-console reset-password           # reset super admin password
 studio-console wizard                   # re-run setup wizard
@@ -190,18 +208,20 @@ studio-console self-update              # upgrade studio-console itself
 studio-console version                  # print version
 studio-console launch-full [--tag T] [--workspace DIR]   # run the Full single-container image on the host
+studio-console launch-core [--tag T] [--workspace DIR]   # run the Core single-container image (external Postgres)
+studio-console core-db-url [URL]        # set Core's external database URL for the next launch-core
 ```
-For the Full image, run operational subcommands through `docker exec studio-full studio-console <cmd>` — see [Run the Full image](#run-the-full-image).
+For the Full or Core image, run operational subcommands through `docker exec studio-full studio-console <cmd>` (or `studio-core`). See [Run the Full image](#run-the-full-image) and [Run the Core image](#run-the-core-image).
 ---
 ## Further reading
-- **[Architecture](docs/architecture.md)** — file layout, state flow, Compose wiring, first-boot, internals (wizard vs init, orphan workers, backup format, major-version detection, deployment contexts).
-- **[Environment variables](docs/env-vars.md)** — every `SHS_*` and supporting var.
-- **[Public hostname topology](docs/topology.md)** — single vs split hostnames, Cloudflare tunnel + Access, IP restrictions.
-- **[VPS + RunPod deployments](docs/vps-runpod.md)** — hybrid setups with GPU worker pods on RunPod.
+- **[Architecture](docs/architecture.md)**: file layout, state flow, Compose wiring, first-boot, internals (wizard vs init, orphan workers, backup format, major-version detection, deployment contexts).
+- **[Environment variables](docs/env-vars.md)**: every `SHS_*` and supporting var.
+- **[Public hostname topology](docs/topology.md)**: single vs split hostnames, Cloudflare tunnel + Access, IP restrictions.
+- **[VPS + RunPod deployments](docs/vps-runpod.md)**: hybrid setups with GPU worker pods on RunPod.
 ---

studio_console-1.3.5/VERSION ADDED Viewed

	@@ -0,0 +1 @@
1	+ 1.3.5

{studio_console-1.3.4 → studio_console-1.3.5}/studio_console/commands.py RENAMED Viewed

@@ -76,7 +76,7 @@ from .wizard import (
 def _apply_scale_flags(up_cmd: list[str], env_data: dict) -> list[str]:
     """Append --scale flags for any *active* worker with a count != 1.
-    Only scales services whose profile is in COMPOSE_PROFILES — compose errors
+    Only scales services whose profile is in COMPOSE_PROFILES, compose errors
     'no such service: X: disabled' if you --scale a profile that isn't active.
     """
     active = {p for p in env_data.get("COMPOSE_PROFILES", "").split(",") if p}
@@ -92,7 +92,7 @@ def _apply_scale_flags(up_cmd: list[str], env_data: dict) -> list[str]:
 def _stop_api_ui_containers() -> None:
     """Remove all studio-api-*, studio-ui-*, studio-nginx-* containers directly.
-    Uses 'ps -a' so Created/Exited containers (not just running ones) are caught —
+    Uses 'ps -a' so Created/Exited containers (not just running ones) are caught
     otherwise a leftover 'studio-api-2' in Created state collides with compose on
     recreate ('container name already in use'). 'docker rm -f' force-removes
     running ones too; '-s' is NOT a valid rm flag.
@@ -167,7 +167,7 @@ def config_menu(context: str, env_file: Path) -> None:
     menu_options = [
         f"Services        {_dim('start · stop · restart · health · logs · links')}",
-        f"Setup           {_dim('wizard — components · secrets · domain · cloudflare')}",
+        f"Setup           {_dim('wizard: components · secrets · domain · cloudflare')}",
         f"Images          {_dim('build · upgrade · rollback')}",
         f"Advanced        {_dim('scale API/UI · per-service ops · cloudflare')}",
         f"Backup          {_dim('backup · restore')}",
@@ -212,7 +212,7 @@ def config_menu(context: str, env_file: Path) -> None:
                 print(f"  Status: {_cyan('configured')}  |  Services → Start all")
             elif mv.scrape_guardrail_failure(env_file, context):
                 print(
-                    f"  Status: {_red('blocked')}  |  major-version boundary — see Services → Health"
+                    f"  Status: {_red('blocked')}  |  major-version boundary, see Services → Health"
                 )
             else:
                 print(f"  Status: {_yellow('starting...')}")
@@ -265,7 +265,7 @@ def _restart_for_setup(
     base = compose_cmd(env_file)
     new_profiles = {p for p in env_data.get("COMPOSE_PROFILES", "").split(",") if p}
-    # Resolve missing images BEFORE stopping anything — a missing image must not
+    # Resolve missing images BEFORE stopping anything, a missing image must not
     # leave a running stack torn down. In source mode we build; in registry mode
     # we pull. Only if a pull fails do we bail, with everything still up.
     missing = _missing_images(env_file)
@@ -279,10 +279,10 @@ def _restart_for_setup(
             if not _pull_images(missing, tag):
                 error(f"Failed to pull images: {', '.join(missing)}")
                 warn("Check the tag (SHS_STUDIO_VERSION) and registry access.")
-                warn("No changes applied — services left as they were.")
+                warn("No changes applied, services left as they were.")
                 return
-    # Stop workers directly — catches orphans whose profiles were removed from .env.
+    # Stop workers directly, catches orphans whose profiles were removed from .env.
     # 'ps -a' so Created/Exited workers are removed too (avoids name collisions on
     # recreate); 'rm -f' force-removes running ones ('-s' is not a valid rm flag).
     _, ps_out = run_quiet(["docker", "ps", "-aq", "--filter", "name=studio-worker-"])
@@ -302,7 +302,7 @@ def _restart_for_setup(
                 timeout=120,
             )
         else:
-            # Workers only — explicit service names so API/UI/nginx are never touched
+            # Workers only, explicit service names so API/UI/nginx are never touched
             worker_services = [
                 svc for svc in SCALE_PROFILES.values() if svc in new_profiles
             ]
@@ -315,7 +315,7 @@ def _restart_for_setup(
                     timeout=120,
                 )
             else:
-                info("No active worker profiles — nothing to start.")
+                info("No active worker profiles, nothing to start.")
         ok("Done")
     except Exception as e:
         error(f"Docker Compose failed: {e}")
@@ -327,13 +327,13 @@ def _restart_for_setup(
 def _submenu_setup(context: str, env_file: Path) -> None:
-    """Setup submenu — runs the wizard then offers Apply now / Skip."""
+    """Setup submenu, runs the wizard then offers Apply now / Skip."""
     old_profiles = (
         read_env(env_file).get("COMPOSE_PROFILES", "") if env_file.exists() else ""
     )
     if wizard(context, env_file):
         apply_options = [
-            f"Apply now  {_dim('restarts changed services — brief downtime possible')}",
+            f"Apply now  {_dim('restarts changed services, brief downtime possible')}",
             f"Skip       {_dim('apply on next manual restart')}",
         ]
         pick = _interactive_single(
@@ -364,7 +364,7 @@ def _cmd_scale_api_ui(env_file: Path) -> None:
     ok("Scaling updated")
     if not _interactive_yn("Restart API + UI to apply?", default=True):
         return
-    warn_header("This will restart API, UI, and nginx — workers will keep running")
+    warn_header("This will restart API, UI, and nginx. Workers will keep running")
     _stop_api_ui_containers()
     env_data = read_env(env_file)
     run(
@@ -377,7 +377,7 @@ def _cmd_scale_api_ui(env_file: Path) -> None:
 def _submenu_services(context: str, env_file: Path) -> None:
-    """Services submenu — daily ops: start/stop/restart all, health, logs, links."""
+    """Services submenu, daily ops: start/stop/restart all, health, logs, links."""
     options = [
         f"Start all      {_dim('pull/build if needed, then start')}",
         f"Stop all       {_dim('stop all containers')}",
@@ -459,7 +459,7 @@ def _owns_postgres(env_data: dict) -> bool:
     """True if SHS_DATABASE_URL points at a postgres instance we manage.
     External DBs (CloudSQL, RDS, etc., or a CloudSQL Auth Proxy sidecar) own
-    their own backup tooling — console refuses backup/restore in those cases
+    their own backup tooling, console refuses backup/restore in those cases
     rather than producing dumps that can't be restored back to the source.
     """
     url = env_data.get("SHS_DATABASE_URL", "")
@@ -511,7 +511,7 @@ def _submenu_backup(context: str, env_file: Path) -> None:
 def cmd_db_role(context: str, env_file: Path) -> None:
     """Show/enable the restricted runtime DB role (shs_app cutover).
-    Console only writes SHS_DATABASE_APP_URL — the API's bootstrap provisions
+    Console only writes SHS_DATABASE_APP_URL, the API's bootstrap provisions
     the role, grants, and RLS posture from it on every boot. SHS_DATABASE_URL
     stays privileged; console's own psql/dump/restore tooling keeps using it.
     """
@@ -527,10 +527,10 @@ def cmd_db_role(context: str, env_file: Path) -> None:
         "SHS_DATABASE_URL", ""
     )
     if not db_url:
-        error("No SHS_DATABASE_URL found — configure the database first.")
+        error("No SHS_DATABASE_URL found. Configure the database first.")
         return
-    warn_header("The API currently connects as the privileged DB role — RLS is inert")
+    warn_header("The API currently connects as the privileged DB role; RLS is inert")
     info("Enabling writes SHS_DATABASE_APP_URL (role shs_app) next to SHS_DATABASE_URL.")
     info("The API provisions the role itself on next boot; console runs no SQL.")
     info(_dim("Requires a studio image with restricted-role support; older images ignore it."))
@@ -542,7 +542,7 @@ def cmd_db_role(context: str, env_file: Path) -> None:
     print()
     if not _interactive_yn("Enable the restricted DB role?", default=False):
-        info("Skipped — no changes made.")
+        info("Skipped, no changes made.")
         return
     set_env_value(env_file, "SHS_DATABASE_APP_URL", derive_app_db_url(db_url))
@@ -560,26 +560,26 @@ def cmd_db_role(context: str, env_file: Path) -> None:
                 ),
                 timeout=120,
             )
-            ok("Services restarted — check Services → Health for the RLS posture.")
+            ok("Services restarted. Check Services → Health for the RLS posture.")
         else:
             info("Takes effect on the next restart of the API.")
     else:
         warn(
-            "Restart this container from the host to apply — provisioning runs in "
+            "Restart this container from the host to apply, provisioning runs in "
             "the container entrypoint, not under supervisord."
         )
         print(f"    {_bold('docker restart <container>')}   {_dim('(or stop/start the pod on RunPod)')}")
 def _submenu_advanced(context: str, env_file: Path) -> None:
-    """Advanced submenu — scale API/UI, per-service ops, Cloudflare ops."""
+    """Advanced submenu, scale API/UI, per-service ops, Cloudflare ops."""
     options = [
         f"Scale API/UI      {_dim('set replica count, enable/disable nginx LB')}",
         f"Start one         {_dim('start a stopped service')}",
         f"Stop one          {_dim('stop a running service')}",
         f"Restart one       {_dim('restart a single service')}",
         f"Show .env         {_dim('current configuration values')}",
-        f"DB role           {_dim('restricted runtime role (RLS) — status · enable')}",
+        f"DB role           {_dim('restricted runtime role (RLS): status · enable')}",
         f"Cloudflare        {_dim('tunnel · routes · IP rules · Access')}",
     ]
     idx = _interactive_single("Advanced", options, default=0)
@@ -759,7 +759,7 @@ def _submenu_cloudflare(context: str, env_file: Path) -> None:
                 _sync_derived_urls(env_file)
                 ok(f"Domain updated to {new_domain.rstrip('/')}")
-                warn("API token not set — DNS records and Access app not updated")
+                warn("API token not set, DNS records and Access app not updated")
             else:
                 warn("No domain entered")
@@ -902,7 +902,7 @@ def cmd_build(env_file: Path, images: list[str] | None, confirm: bool = True) ->
         running = [line.strip() for line in out.strip().splitlines() if line.strip()]
         if running:
             warn_header(
-                "Restart required to apply new images — Studio will be briefly unavailable"
+                "Restart required to apply new images. Studio will be briefly unavailable"
             )
             info(f"Running services: {', '.join(running)}")
             if _interactive_yn("Restart now?", default=True, nav=False):
@@ -963,7 +963,7 @@ def cmd_start(context: str, env_file: Path) -> None:
                     run(compose_cmd(env_file) + ["pull"], timeout=600)
                 except Exception:
                     error("The download was interrupted before it finished.")
-                    warn("Run Start again — it should pick up where it left off.")
+                    warn("Run Start again, it should pick up where it left off.")
                     return False
                 ok("Images pulled")
@@ -1077,7 +1077,7 @@ def _core_plan(
     container: str, pg_container: str, nginx_port: int | str = 80
 ) -> "_BootstrapPlan":
     # Core's API runs in `container`, but Postgres is an external sidecar
-    # (`pg_container`) — so password hashing execs into the API container while
+    # (`pg_container`), so password hashing execs into the API container while
     # every psql runs against the sidecar. base omits "exec"; helpers append it.
     # Core launches with publish_internal=False, so port 8000 is not published:
     # reach the API through the front door.
@@ -1134,7 +1134,7 @@ def _unset_env_for_bootstrap(
 def _super_admin_exists(plan: "_BootstrapPlan", env_data: dict) -> bool:
-    """True if a super_admin user is already in the DB — the bootstrap gate."""
+    """True if a super_admin user is already in the DB, the bootstrap gate."""
     pg_user = env_data.get("POSTGRES_USER", "postgres")
     rc, out = run_quiet(
         plan.base
@@ -1251,7 +1251,7 @@ def _bootstrap_first_admin(
             plan,
         )
         return True
-    warn("Bootstrap incomplete — will retry account creation on next start.")
+    warn("Bootstrap incomplete, will retry account creation on next start.")
     return False
@@ -1574,7 +1574,7 @@ def _app_db_url(env_data: dict) -> str:
 def _print_db_role_posture(env_data: dict, api_up: bool) -> None:
     """One-line RLS posture. A healthy API with the app URL set proves
-    restricted mode — boot is fail-closed on an RLS-inert role."""
+    restricted mode, boot is fail-closed on an RLS-inert role."""
     if _app_db_url(env_data):
         if api_up:
             print(
@@ -1584,12 +1584,12 @@ def _print_db_role_posture(env_data: dict, api_up: bool) -> None:
         else:
             print(
                 f"  {'DB role':24s} {_yellow('restricted (configured)')}  "
-                f"{_dim('API down — an RlsInertError in its logs means SHS_DATABASE_APP_URL is misconfigured')}"
+                f"{_dim('API down, an RlsInertError in its logs means SHS_DATABASE_APP_URL is misconfigured')}"
             )
     else:
         print(
             f"  {'DB role':24s} {_yellow('privileged')}  "
-            f"{_dim('RLS inert, app-layer checks only — enable via the DB role menu')}"
+            f"{_dim('RLS inert, app-layer checks only, enable via the DB role menu')}"
         )
@@ -1864,7 +1864,7 @@ def cmd_reset_password(context: str, env_file: Path) -> None:
         info("Resetting admin password...")
         try:
             # SHS_ names are the app script's env contract (scripts/reset_admin_password.py
-            # reads SHS_ADMIN_PASSWORD / SHS_FORCE_PRODUCTION) — keep them, do NOT rename
+            # reads SHS_ADMIN_PASSWORD / SHS_FORCE_PRODUCTION), keep them, do NOT rename
             # to CONSOLE_. Transient injection only; never written to .env.
             run(
                 compose_cmd(env_file)
@@ -1884,7 +1884,7 @@ def cmd_reset_password(context: str, env_file: Path) -> None:
             error("Failed to reset password. Check that the API container is healthy.")
     else:
         info("Resetting admin password...")
-        # SHS_ names are the app script's env contract — keep, do NOT rename to CONSOLE_.
+        # SHS_ names are the app script's env contract, keep, do NOT rename to CONSOLE_.
         os.environ["SHS_ADMIN_PASSWORD"] = new_password
         os.environ["SHS_FORCE_PRODUCTION"] = "true"
         try:
@@ -2039,7 +2039,7 @@ def _format_local_time(iso_string: str) -> tuple[str, str]:
 def _verify_post_restore(context: str, env_file: Path, db_file: str) -> None:
     """After restore, confirm the live DB's alembic_version matches the dump's.
-    Mismatch means the apply succeeded but landed on the wrong schema — surface
+    Mismatch means the apply succeeded but landed on the wrong schema, surface
     it loudly so the operator doesn't trust a silent success.
     """
     expected = _read_revision_from_dump(db_file)
@@ -2111,7 +2111,7 @@ def _restore_preflight(context: str, env_file: Path, db_file: str) -> bool:
     # Encryption-key check: the dump's encrypted columns are only readable with
     # the key they were written under. We never store that key in the backup, so
-    # we compare fingerprints — live key vs. the fp recorded in the header.
+    # we compare fingerprints, live key vs. the fp recorded in the header.
     backup_key_fp = header.get("encryption_key_fp", "")
     live_key_fp = _encryption_key_fingerprint(env_data)
     key_mismatch = bool(backup_key_fp and live_key_fp and backup_key_fp != live_key_fp)
@@ -2172,7 +2172,7 @@ def _restore_preflight(context: str, env_file: Path, db_file: str) -> bool:
         )
         if _offer_key_recovery(env_file, backup_key_fp):
             return True  # key now matches; fall through to restore
-        info(_cyan("Restore needs the matching key — the data is unreadable without it."))
+        info(_cyan("Restore needs the matching key. The data is unreadable without it."))
         info("Aborted (no valid encryption key).")
         return False
@@ -2259,7 +2259,7 @@ def cmd_backup(
     ok(f"Backup complete: {backup_dir}")
-    # The encryption key is deliberately NOT stored in the backup — co-locating
+    # The encryption key is deliberately NOT stored in the backup, co-locating
     # it with the ciphertext would let anyone holding the backup decrypt it.
     # Skip the notice for internal pre-restore snapshots (not operator-facing).
     if what in ("all", "db") and name_prefix != "pre-restore":
@@ -2275,7 +2275,7 @@ def _print_encryption_key_notice(env_data: dict) -> None:
     print(_bold("⚠ Record your encryption key separately"))
     print("─" * 40)
     if key:
-        print("This backup does NOT contain your encryption key — by design, so a")
+        print("This backup does NOT contain your encryption key, by design, so a")
         print("stolen backup can't be decrypted. The dump is unrecoverable without it.")
         print()
         print("Copy SHS_CREDENTIAL_ENCRYPTION_KEY from your .env into a password manager or")
@@ -2283,7 +2283,7 @@ def _print_encryption_key_notice(env_data: dict) -> None:
         print()
         print(f"    SHS_CREDENTIAL_ENCRYPTION_KEY={key}")
     else:
-        print(_yellow("No SHS_CREDENTIAL_ENCRYPTION_KEY is set in .env — nothing to record."))
+        print(_yellow("No SHS_CREDENTIAL_ENCRYPTION_KEY is set in .env. Nothing to record."))
     print()
@@ -2416,7 +2416,7 @@ def cmd_restore_db(
     try:
         if context == "host":
-            # Stop API/UI before dropping the schema — a live API reconnects
+            # Stop API/UI before dropping the schema, a live API reconnects
             # to postgres mid-restore (pg_terminate_backend only kills current
             # connections) and reads/writes a half-restored schema. The
             # pre-restore snapshot above is the recovery net if this fails.
@@ -2512,10 +2512,10 @@ def cmd_restore_db(
     ok("Database restored")
     _verify_post_restore(context, env_file, db_file)
-    # .env (and its encryption key) is intentionally left untouched — the key
+    # .env (and its encryption key) is intentionally left untouched, the key
     # is never stored in the backup. The preflight already gated key mismatch.
-    # API/UI were stopped for exclusive DB access — bring them back up.
+    # API/UI were stopped for exclusive DB access, bring them back up.
     if context == "host":
         info("Restarting API/UI...")
         env_data2 = read_env(env_file)
@@ -2533,14 +2533,14 @@ def cmd_restore_db(
             )
     else:
         # Provisioning/grants/RLS re-apply in the container ENTRYPOINT, so the
-        # restore flow ends with a container restart — supervisorctl restart
+        # restore flow ends with a container restart, supervisorctl restart
         # is not enough. Until then the restricted role authenticates but every
         # table access is permission-denied → sanitized 500s (fail-closed).
         warn("Restart this container from the host now to finish the restore.")
         print(f"    {_bold('docker restart <container>')}   {_dim('(or stop/start the pod on RunPod)')}")
         if _app_db_url(env_data):
             warn(
-                "Until the restart, API requests fail with sanitized 500s — "
+                "Until the restart, API requests fail with sanitized 500s, "
                 "the restore dropped the restricted role's grants; boot re-applies them."
             )
@@ -2579,7 +2579,7 @@ def cmd_restore(
             db_name = "selfhost_studio"
             if context == "host":
-                # Stop API/UI before dropping the schema — see cmd_restore_db.
+                # Stop API/UI before dropping the schema, see cmd_restore_db.
                 # A live API reconnects mid-restore and reads a half-restored
                 # schema; the pre-restore snapshot above is the recovery net.
                 info("Stopping API/UI for exclusive DB access...")
@@ -2668,7 +2668,7 @@ def cmd_restore(
                     )
             ok("Database restored")
             _verify_post_restore(context, env_file, db_file)
-            # .env (and its encryption key) is intentionally left untouched — the
+            # .env (and its encryption key) is intentionally left untouched, the
             # key is never stored in the backup. Preflight gated key mismatch.
         else:
             warn(f"No database.sql in {restore_dir}")
@@ -2684,7 +2684,7 @@ def cmd_restore(
         else:
             warn(f"No orgs.tar.gz in {restore_dir}")
-    # API/UI were stopped for exclusive DB access — bring them back up.
+    # API/UI were stopped for exclusive DB access, bring them back up.
     if context == "host":
         print()
         info("Restarting API/UI...")
@@ -2812,7 +2812,7 @@ def cmd_upgrade(context: str, env_file: Path) -> None:
     run(compose_cmd(env_file) + ["pull"], timeout=300)
     ok("Images pulled")
-    # Restart — reapply scale flags + --remove-orphans so replica counts survive
+    # Restart, reapply scale flags + --remove-orphans so replica counts survive
     # the upgrade and stale containers from the old version are cleaned up.
     info("Restarting services...")
     env_data = read_env(env_file)
@@ -2870,7 +2870,7 @@ def _detect_install_method() -> str:
     if marker.exists():
         return marker.read_text().strip()
-    # uv tool install — path contains uv/tools, no pip module available
+    # uv tool install, path contains uv/tools, no pip module available
     if "uv/tools" in str(pkg_dir).replace("\\", "/"):
         return "uv"
@@ -2882,7 +2882,7 @@ def _detect_install_method() -> str:
     if "Cellar" in str(pkg_dir) or "homebrew" in str(pkg_dir).lower():
         return "brew"
-    # Inside a git repo — dev mode
+    # Inside a git repo, dev mode
     if (pkg_dir / ".git").exists() or (pkg_dir.parent / ".git").exists():
         return "dev"
@@ -2890,7 +2890,7 @@ def _detect_install_method() -> str:
 def cmd_self_update(context: str) -> None:
-    """Update studio-console — delegates to the correct mechanism for the install method."""
+    """Update studio-console, delegates to the correct mechanism for the install method."""
     import json
     import os
     import shutil
@@ -2905,14 +2905,14 @@ def cmd_self_update(context: str) -> None:
     if context in ("container", "runpod"):
         warn("studio-console is baked into the container image.")
-        warn("Update by pulling a new image tag — not via self-update.")
+        warn("Update by pulling a new image tag, not via self-update.")
         return
     method = _detect_install_method()
     info(f"Current version: {__version__}  (installed via {method})")
     if method == "dev":
-        warn("Running from source — use git pull to update.")
+        warn("Running from source. Use git pull to update.")
         return
     if method == "brew":
@@ -2927,10 +2927,16 @@ def cmd_self_update(context: str) -> None:
             [os.sys.executable, "-m", "pip", "install", "--upgrade", "studio-console"],
             timeout=120,
         )
-        ok("Updated — restart studio-console to use the new version")
+        ok("Updated. Restart studio-console to use the new version.")
         return
-    # uv and curl installs both resolve the latest GitHub release first
+    if method == "uv":
+        info("Updating via uv tool...")
+        run(["uv", "tool", "install", "--force", "studio-console"], timeout=120)
+        ok("Updated. Restart studio-console to use the new version.")
+        return
+    # curl installs resolve the latest GitHub release and extract its tarball
     info("Checking for updates...")
     api_url = "https://api.github.com/repos/selfhosthub/studio-console/releases/latest"
     try:
@@ -2953,16 +2959,6 @@ def cmd_self_update(context: str) -> None:
     info(f"Updating {__version__} → {latest_version}")
-    if method == "uv":
-        wheel_url = (
-            f"https://github.com/selfhosthub/studio-console/releases/download/"
-            f"{latest_tag}/studio_console-{latest_version}-py3-none-any.whl"
-        )
-        info("Updating via uv tool...")
-        run(["uv", "tool", "install", "--force", wheel_url], timeout=120)
-        ok("Updated — restart studio-console to use the new version")
-        return
     tarball_url = f"https://github.com/selfhosthub/studio-console/archive/refs/tags/{latest_tag}.tar.gz"
     install_dir = Path(__file__).resolve().parent.parent