LookoutSentry
It checks every connection, setting and domain your systems depend on, and tells you what broke and how to fix it, without touching anything.
Built on
Where it lands
You don't need
Slack or Notion. It lands where you already work.
01The problem
A key goes missing, a domain record never gets added, a database never gets created. Nothing shouts. You find out when an order doesn't arrive or a customer says the site is down.
Without it
Found by accident. A missing setting turns up during a manual audit, weeks later.
"Connected" means nothing. A key is pasted, so everyone assumes it works.
A domain that serves nothing. Records are missing and nobody looks.
An error with no next step. You know something failed, not what to do.
With Lookout
Checked every day. Every setting each system needs is checked on a schedule.
Connected means tested. Credential present, handler built, and a live check passed. Anything less is named.
Domains watched. The site's DNS record checked on Cloudflare, and the site itself checked for status, speed and certificate.
A fix card per failure. What failed, where, and the step most likely to fix it.
02How it works
Every run follows the same five steps. Nothing it writes goes out until a named person approves that exact draft: Tony on our own brands, and on yours, the person you name at setup.
From each brand's settings file: its connections, the keys they need, its website and its scheduled engines, plus an optional watch file naming extra pages, the Vercel project and the GitHub repos.
Read only. A live check on every connection, each key present or missing (never read), the site's status, speed and certificate, the latest Vercel deploys, n8n's health check, the approval queue, whether each engine ran on time, the site's DNS record on Cloudflare, and the latest commit and workflow run on GitHub.
Critical, warning or info, or "unknown" with the reason when a check could not run. A check that cannot run is never reported as fine and never as down.
Each new failure gets two plain sentences, what it means and the likely fix, written by Claude from the finding alone. When a check runs without AI, it gets a fixed sentence instead.
A Telegram message to you only when something new is critical. The fix is yours to make; it never changes a setting itself.
03Sample output
What Lookout produces, shown on labelled sample data. Names and figures are fictional; your own data replaces them.
Connections
04What you get
Installed on your accounts, tested on your real data with every message held as a draft, then hosted, watched and fixed by us.
Scheduled checkson every brand's connections, keys, website, DNS record, Vercel deploys, n8n, GitHub repos, the approval queue and engine runs.
Four honest statusesper connection: connected, needs credentials, not built, error. Connected only after a live check passes.
Fix cardsWhat failed and the likely fix, in two plain sentences, one per new failure.
AlertsNew critical failures listed in the Vanguard, and sent to your phone through Telegram once the bot is connected.
Repeats groupedThe same failure tomorrow goes under yesterday's card, not on top of it.
Feeds the health gridon the Vanguard, from the findings file each run writes.
A plain English guide and a walkthrough video
Fixes and updates included for as long as you're on the care plan
05Who it's for
It fits if
Not for you if
Your whole business runs on one hosted tool with no keys or connections of your own. Want the failures read to you each morning with everything else? That's the Daybreak Brief.
Proof
Built for Sod-X first. No numbers yet. We publish them after 30 days of real runs, each one linked to the run that produced it.
06Install Lookout
One setup, then one monthly line. The final number comes in writing on a Project Quote before anything is built.
Setup · Starter build
$3,500 to $6,000
One time. We install it, connect it to your accounts and make it yours.
Then monthly · Maintain care plan
$750/mo
We host it, watch it and fix it, so you never have to.
07Questions
The Maintain care plan, $750/mo, covers hosting, monitoring and fixes to the watchdog itself. AI usage for writing the fix cards is billed at cost on its own line. Cloudflare, GitHub and Telegram stay on your own plans.
About 14 days. It connects to your DNS, code and alert accounts per brand, and each check is tested against a known good and a known broken case before it goes live.
Read-only tokens for your hosting (Vercel), DNS (Cloudflare) and code (GitHub) accounts, the pages and repos to watch, and a list of what each brand is supposed to have switched on.
It never changes a setting, pastes a key, edits a DNS record, restarts a service, re-runs a workflow or deploys, and it messages no one but you. It reads and reports. The fix is yours to make, or ours on your say.
Yes, where a brand runs n8n: it calls the instance's health check each run, and a bad answer, confirmed by a second try, is critical. It does not look inside individual workflows.
08For agencies and operators
If you run more than one business, or you serve clients who need the same thing, we install any engine in this catalog for each of them, then run it and fix it behind you.
Partner terms, including whether the client screens carry your name and logo, are agreed in writing on the partner call.
09Works well with
One plain English brief each morning: what moved overnight, what is stuck, and what needs your yes.
A password protected screen per brand showing every engine, its last run, its connectors and every draft waiting on you.
Next step
Thirty minutes on a call. Bring the job that eats your week, and we will tell you whether Lookout fits, or that nothing in the store does yet. If it fits, the number comes in writing.
