Stand up a server, connect it end to end, and clone a whole server to a new one — all from the terminal — plus the first privileged writes the SpinupWP API can't do.
Highlights
- Clone a server (
C). A guided wizard clones one or more of a server's sites onto a new/existing destination — plan + size, provision/pick the dest, pull each site server-to-server (Standard WP and Bedrock), verify against the source, and repoint DNS only when you say so. Runs in the background.
- Create & connect servers (
c / V). Provision a server priced from the provider catalog (DigitalOcean / Vultr / Linode / Hetzner), then connect a 0-site box end to end with a vanity site (DNS → site → HTTPS → SSH-key handoff).
- Privileged writes over SSH: sudo (
S) + SSH keys (K). Connect sudo for the session, then grant/revoke your personal key and/or Spinup's dedicated machine key — one site or every site on the server.
- Save your sudo password to the macOS Keychain (opt-in). Auto-unlocks sudo next session; the password lives only in the Keychain + memory, never
config.json.
- Completion toasts. PHP upgrades and server reboots/restarts raise a toast when they finish.
Notes
- The SSH actions (sudo, key grant, clone pull) use your own SSH access, not the API token. Create / clone / PHP / reboot need a Read/Write token.
- Clone never touches DNS until the final cutover step, which you trigger.
- The Keychain save is opt-in and macOS-only.
Local-dev images that just work, plus a version/About panel and self-announcing updates. This release also renames the app to Spinup.
Highlights
- Production media fallback (
m). After a p DB pull, your local site shows broken images because the media library isn't synced. Press m on a linked WordPress site to serve any missing-locally upload straight from production — no file copying. It runs in WordPress (pure PHP), so it works the same on Valet / Herd / LocalWP / DDEV / MAMP with no web-server config, and it covers page-builder URLs (Elementor), legacy /wp-content/uploads paths from a Bedrock conversion, and CDN/S3 redirects. Local-only, read-only on production, and self-disabling on the production domain.
- Version & About. The header shows
◆ Spinup vX.Y.Z, and the ? help overlay is redesigned into a clean multi-column layout with an About column (version, how to update, repo).
- Update notifications. On launch (checked at most once every 6h), Spinup compares your version to the latest release and shows a
✦ vX.Y.Z hint when there's a newer one — so future updates announce themselves.
Notes
- This is the release that introduces the update check — so the nudge starts working once you're on v0.8.0; from here on, new releases announce themselves. Updates stay developer-centric (
git pull).
- The app is now Spinup (a control center for your SpinupWP account). References to the SpinupWP service — the
w web app, the API, deep links — are unchanged.
- The media fallback's only write is a local mu-plugin in your linked copy; production stays read-only.
Quality-of-life polish on the database backup (d) and prod→local sync (p) overlays.
Highlights
- Progress now reads as a building checklist. Both
d and p show every stage (back up local, export, download, import, rewrite URLs, run hook) in one bordered panel — each gets a ✓ as it finishes, the running stage spins, and the saved paths/size appear below the completed stack. The old separate "running" and "Done" screens are now a single frame, so you see the whole operation end to end.
- Failures point at the exact step. If a stage breaks, that row is marked
✕ with the error inline — no guessing which part failed.
- Scaffold a post-import hook in place. Press
p on a linked site with no bin/sync.d/post-import.sh and the confirm screen explains what the hook is and offers s to write an inert sample — documented with the WEB_DIR / SYNC_REMOTE_HOST / SYNC_LOCAL_HOST env contract and commented examples (Elementor URL swaps, plugin toggles), so nothing runs until you uncomment it.
Notes
- No new SpinupWP API calls; the only new write is a local file (the sample hook), and only when none exists. Production stays read-only.
- Progress still survives closing the overlay — reopen with the same key to watch it through.
The first DNS write, plus a one-keystroke way to pull a production database into your local copy — all from the terminal.
Highlights
- Edit DNS record TTLs — the first DNS write. From the DNS migration view,
⏎ on a site's hosting record opens a focused editor to change its TTL (preset or custom), written to AWS Route 53 or Cloudflare behind a confirm step. The prep step for a low-risk cutover: lower the TTL, repoint, restore. An edit-time check re-reads the live nameservers and blocks the write if your account's zone isn't the one actually serving the domain. Cloudflare proxied records are read-only.
- DNS view is now a migration lens. Organized by site (labeled by its own domain, even subdomains), each with its hosting record's type / TTL (in seconds) / value, a
◀ here flag when it points at this server, and a +www tag when www follows the apex. n scopes to one site; a expands to the whole server.
- Download a production DB backup (
d). On a linked WordPress site, export the production database with wp-cli, gzip it, and download it into the copy's sql/ folder — read-only on production, no local WP-CLI needed.
- Pull production → local (
p, opt-in). Refresh your local database from production: backs up local first, imports, rewrites URLs (wp search-replace), and runs your existing bin/sync.d/post-import.sh hook. Works with Standard WP and Bedrock with no per-site config. It overwrites your local DB, so it's off by default — enable with localSync.
- Site Control panel. The Search action menu is now grouped (Open / Remote / Local / Server) so the growing per-site action set stays scannable.
Notes
- DNS edits only ever touch a site's own hosting records — your MX, TXT, and other zone records are never shown or modified.
- Editing a Cloudflare record needs a token scoped to
Zone.DNS:Edit (the read-only Zone:Read token isn't enough).
- The DB backup/sync actions appear only for WordPress sites that are linked to a local copy; the destructive sync additionally requires the
localSync opt-in.
See where every domain's DNS is hosted — and whether you can edit it — without leaving the terminal. A new read-only DNS module; no DNS records are ever changed.
Highlights
- DNS host inventory.
n on a site, or N on a server, for a zone-keyed map of where every domain's DNS is hosted (Cloudflare, AWS Route 53, GoDaddy, …) — the inventory you'd otherwise build by hand to migrate or clone a site. www + apex collapse to one zone; separate-TLD redirects surface as their own.
- Access detection. Each zone shows whether you can edit it —
✓ editable · ↗ web only · ○ needs key · · unknown — checked against the account that actually serves the zone (its live nameservers), so a stale or duplicate zone never shows a false green. An ACCOUNT column names the owning account when more than one is connected.
- Connect providers (
c). AWS Route 53, Cloudflare, and GoDaddy — multiple accounts each, verified on the spot, secrets masked and stored locally (env vars honored).
- GoDaddy web fallback (
w). Opens your GoDaddy Clients hub and copies the domain to your clipboard, for the common Delegate-Access workflow.
Notes
- The DNS module is entirely read-only; provider credentials are optional (without them you still get the full host inventory).
- API tokens and provider secrets are now masked as you type, including on the first-run onboarding screen.
Bridge SpinupWP's remote inventory to your local dev copies — link, discover, and act on them — plus a one-key SSH. All read/local-only; no new API writes.
Highlights
- Link a local copy — press
L on a site (Servers / Stacks / Search) to link its local checkout (path + local URL). Details gain a "Local" field; open it with t (terminal at the path) or v (local URL in the browser).
- Auto-discover — press
S in Stacks to scan your folders and match subdirectories to sites by git remote → WP_HOME → folder name, then batch-link. Add scan folders inline (comma-separated) the first time.
- "Needs a local copy" report — press
f in Stacks for the managed sites you have no local copy for, alphabetical and filterable by stack (←/→); a trims to the ones needing attention.
- Drift at a glance — a linked, on-disk site shows
⇡N unpushed / ● uncommitted in the context strip (read locally, no network).
- SSH into a site — press
s to open a new terminal already running ssh into it.
- Glanceable row markers —
◆ linked locally, ↑N pending WordPress updates, across every site list.
- "Explain this screen" — press
i on any view for a plain-language guide; plus per-view subtitles under the tab bar.
Notes
- Terminal actions open your terminal (detected via
$TERM_PROGRAM, e.g. iTerm; override with terminalApp), not a hardcoded Terminal.app. SSH / "open a terminal" script iTerm or Terminal on macOS; other terminals fall back to Terminal.app.
- The local URL is tool-agnostic (Valet, Cove, LocalWP, Herd, DDEV, …); an older
valetUrl config key is migrated automatically.
- Outcome-oriented relabels: "detect" → "identify app", "scan local" → "find local copies", "upgrade PHP" → "change PHP".
The second write action: reboot a server or restart a service, right from the dashboard.
Highlights
- Server actions — press
a on a server (Servers tab, or a result in Search) to reboot it (POST /servers/{id}/reboot) or restart a service: Nginx / PHP-FPM / MySQL / Redis (POST /servers/{id}/services/{svc}/restart). Pick → confirm → the event is tracked to completion. Like PHP upgrades, it runs in the background: close the overlay with Esc and the server's row keeps a spinner until it settles.
- Reboot is the careful one — its confirmation calls out that it takes the whole server down briefly (every site on it); a service restart is a brief blip.
- Reboot visibility — servers needing a reboot show a
↻rbt badge in the Servers list and in Search results (and were already on the Dashboard's "Needs attention" panel).
- Why a reboot is pending — the SpinupWP API exposes only a
reboot_required boolean with no reason, so the overlay reads Ubuntu's /var/run/reboot-required (+ .pkgs) over SSH (read-only, reusing the health-view connection) and shows the pending packages — typically a kernel/security update. Labeled as OS-level context, not as SpinupWP's internal logic (a fleet-wide check confirmed the boolean tracks that file 1:1).
Notes
- These actions need a SpinupWP Read/Write token; with a Read Only token they return a clear "token is read-only" message and nothing changes.
- Reuses the write scaffolding introduced in v0.3.0 (mutate / event polling / confirm overlay).
The app's first write action: upgrade a site's PHP version — and Search results become actionable.
Highlights
- Upgrade a site's PHP version — press
u on a site (Servers / Stacks / Search) to pick a version and apply it (PUT /sites/{id}/php). The upgrade event is tracked to completion.
- Dynamic version picker — the offered versions come from the live endoflife.date PHP schedule (refreshed + cached), so new releases like 8.5 appear on their own; the current version is marked and EOL versions are flagged. Versions aren't filtered to "installed" — SpinupWP installs one on demand the first time a site uses it, so the full supported set is always offered.
- Runs in the background — tracking lives in the app's store, so you can press
Esc to close the modal and the upgrade keeps going. The site's row shows a spinner and the target version (→8.3) until it settles, then refreshes to the new version (or flags ⬆! on failure). Several sites can upgrade at once; the detail panel mirrors the in-progress state.
- Search becomes a command surface — press
Tab (or →) to hand focus from the search box to the selected result's action menu: o open · w SpinupWP · u PHP upgrade · h health. ← / Esc returns to typing.
- Read-only-safe — writes need a Read/Write token. With a Read Only token an upgrade returns a clear "your token is read-only" message and nothing changes; the entire read path keeps working.
Notes
- A site whose server has a pending SpinupWP platform upgrade can't be managed via the API until that runs — the picker is blocked and points you to the
w web deep link.
- This is the first write feature: the shared scaffolding (write-capable client, async event polling, confirm-before-firing overlay) is now in place to build the next write actions on.
Know what's actually running across your fleet. This release adds site stack detection, a new Stacks tab, and on-demand SSH probing that names what the SpinupWP API can't.
Highlights
- New
3 Stacks tab — fleet composition at a glance: counts and bars for Standard WP / Bedrock / Non-WP, drill-down into the sites of any bucket, and the fleet-wide PHP version distribution with end-of-life versions flagged.
- Tier-1 classification (instant, no SSH) — every site is bucketed from data already fetched. Stack tags also appear in the Servers and Search lists, and a Stack field in the site detail panel.
- Tier-2 SSH probe — press
d to inspect a site's filesystem (read-only) and identify it precisely: WordPress (with version), Bedrock, WHMCS, Laravel, or Static HTML. Press D to probe an entire stack in one go (bounded concurrency).
- Catches API mislabels — WordPress installed outside SpinupWP's installer reports
is_wordpress=false; a conclusive probe overrides that, moving the site into its true bucket. The Non-WP bucket expands into named sub-rows (WHMCS / Laravel / Static HTML / Unknown / unprobed).
- Cached & persistent — probe results are written to
~/.config/spinupwp-tui/stack-cache.json, hydrated at startup, so detections survive restarts without re-running SSH.
Fixed
- Capital-letter shortcuts (
D, and the previously broken G jump-to-bottom) now register correctly.
Notes
- Still read-only against the SpinupWP API; the SSH probe runs only read-only filesystem checks, using your existing SSH access.
- Tab keys shifted: Search is now
4, Events is 5.