For founders with side projects
You moved on. I kept watching.
Know which of your projects needs you. Paste one URL for a first look at its page response and domain signals. No account needed.
One page load, no clicks or sign-in.
- Seven days on me, no card
- No install
- Hosted in the EU
An illustrative video of example projects. Five side projects, and I open each one the way a visitor does. When the Fit Meal Plan checkout breaks at 10:02, I stay quiet and check again ten minutes later. It breaks again, so I write, with both screenshots. The broken project moves to the top of your list, and the other four still work.
In one minute
See how I watch one project.
Paste one URL, show me the journey once, and I write only when the same check breaks twice.
An illustrative walkthrough of an example project. You paste one URL and see the first result in seconds, without an account. You show me the journey once, and I replay it daily and after detected deploys, within your plan’s daily allowance. I write only when the same check breaks twice, with the evidence I captured. I show the speed of your checks before and after detected deploys, never as an alert. I never call a project fine unless I checked it.
First look
Useful evidence before you connect.
Paste one URL for a first look at its page response and domain signals. No account needed. One page load, no clicks or sign-in.
Page response
Does the address answer, and what came back?
Example: 200 OK · answered in 0.1 s · no redirects
Certificate and domain
A separate lookup for the signals behind the address.
Example: certificate · 72 more days · Let’s Encrypt
Checks worth running
I read your page and name up to three checks worth keeping an eye on.
Example: sign-up · checkout · contact
See a real one
Three snapshots, three honest answers.
The same page you get for your own site. I say when everything I looked at was working, when something deserves a look, and when I simply couldn’t see.
One list. Four honest states.
When a project breaks, it comes first.
All your projects stay in one list. If the same check breaks twice, I move that project to the top and show you what needs attention.
I never call a project fine unless I checked it. Each row shows what I checked, what changed, and when I last looked.
- still
Everything I checked worked.
- broken
The same check broke twice, so it needs you.
- healed
The page changed; I found the same thing in its new place.
- unknown
I couldn’t verify it, so I won’t call it fine.
Watch recorder
Show me the journey once. I’ll replay it daily and after detected deploys.
- Walk it onceClick through the route you want me to repeat. Each click becomes a step while you go.
- Shape the checkRemove or reorder steps, add conditions, then try the whole check before you save it.
- I take it from hereAfter detected deploys and on the daily schedule, within your plan’s daily allowance. Only when it breaks twice in a row do I write.
I record the steps you show me, not your visitors. The recorder and its checks stay outside your app.
After you connect
You build. I take the late shift.
For ongoing checks, add a project and choose or record the steps you want repeated.
I find a useful place to start
I open the homepage and look for what’s worth checking. I try every suggestion once and keep only the ones that work. You review what I kept or record your own.
I look after deploys and every day
Once deploys are connected, I run your checks daily and after detected deploys, within your plan’s daily allowance. A daily run covers the days when you ship nothing.
I check twice before I write
One bad run stays quiet. I run the same check again 10 minutes later. I tell you a check is broken only after it breaks twice in a row, with the evidence I captured. Errors on my side do not count.
Why the first bad run stays quiet
- Deploydetected
- brokestays quiet
- broke againsame check
- I writewith the evidence
- works againI tell you too
- Daily runcovers days you ship nothing
Read how deploys start runs and why the first bad run stays quiet.
From evidence to your next deploy
A useful handoff when a check breaks.
I show you what happened. You take the evidence to your coding assistant, review the repair and deploy it. A later run shows whether the same check works again.
Illustrative example · Fit Meal Plan checkout. These are example events, not a result from your project.
Read the evidence
The checkout shows “Something went wrong. Please try again.” instead of “Thank you for your order”. Open the broken run to see the captured evidence and the step where it stopped.
Copy for your AI
In your project, use Copy for your AI on a supported broken run. I copy a prompt built from that run’s stored facts and a link back to it. Paste it into your coding assistant.
Review the repair and deploy
Your coding assistant helps you make the change. You review it and deploy it. I do not repair or deploy your application for you.
Wait for the next verified result
After the repair, a later run of the same checkout reaches “Thank you for your order”. That run is the evidence that it works again; copying the prompt or sending a deploy signal is not.
Speed around deploys
How long a check takes is not an alert.
I keep how long each check took, from the runs I make daily and after detected deploys, within your plan’s daily allowance. A check that takes longer has still passed, and I never write about it.
For a deploy I noticed, I show the same numbers before and after, side by side, with no verdict and no cause.
With the optional snippet I also see how quickly pages load for the people using your application, and I write once when they take longer after a detected deploy.
Read where I put the numbers and how visitor speed works.
Clear boundaries
What I won’t do
Useful watching needs clear boundaries. These are mine.
I won’t ask for access to your systems
I read your application through a browser, from the outside. No database access, no agent on your servers. If I disappeared tonight, your application would keep working.
I won’t call a check broken after one bad run
I retry that check first and call it broken only when the same check breaks twice in a row. If a run broke on my side, it never counts against you.
I won’t keep recordings past 30 days
On every plan. Recordings may contain names or email addresses, so longer history keeps what happened, not the video of it.
I won’t let AI decide what’s broken
AI only proposes checks and, as a last resort, finds an element that moved. What worked or broke is always decided by the run, not by a model.
“I used to click through the same steps after every deploy. I wanted to show that journey once and have it checked again without having to remember to do it myself.”
Pricing
A fair price for quieter nights.
Seven days on me. Then from $12 a month. No card.
Small
Every trial starts hereOne or two side projects and the checks that pay.
$12a month
$120a year
- Projects
- 3
- Active checks per project
- 3
- Alert destinations
- 3
- Deploy runs a day
- 20
- Run history
- 30 days
Growing
More projects, more checks each, three months of history.
$24a month
$240a year
- Projects
- 10
- Active checks per project
- 5
- Alert destinations
- 5
- Deploy runs a day
- 20
- Run history
- 90 days
Portfolio
Many projects that ship often, and a year of history.
$59a month
$590a year
- Projects
- 30
- Active checks per project
- 7
- Alert destinations
- 10
- Deploy runs a day
- 100
- Run history
- 12 months
On every plan
- Daily checks and checks after detected deploys, within your plan's daily allowance
- Run recordings kept for 30 days
- Certificate and domain checks
- Findings on day one
- Deploy timeline
- Speed of your checks before and after detected deploys, never an alert
- Watch recorder
- Healed checks: I find a moved element again
- Visitor errors with an optional snippet
- Email, Slack, Discord or your own webhook
- No usage charges
Deploy runs share one daily allowance across your account. How deploy runs are counted, refunds and billing
Questions worth asking
What happens after the first look?
Do I have to install anything?
No. Start with a URL. The first look and outside checks need no install. An optional deploy signal helps me notice a deploy sooner. Visitor errors are separate and need an optional one-line snippet.
Will one bad run wake me up?
No. When a check breaks, I doubt it and run the same check again ten minutes later. I open an incident and write only if it breaks twice in a row.
What can I check before paying?
Paste one URL on this page and see the first result. If you ask me to keep watching, you get seven days without entering a card.
Do you need my database or hosting login?
No. I never need a database key to watch a project. Checks run outside your systems and only use sign-in details when a check genuinely requires them.
How long do you keep screenshots and video?
Thirty days, and no longer. They can contain names or email addresses from your app, so I would rather not keep them indefinitely.
One URL. One useful answer.
You don’t have to keep staring at it.
It costs you nothing to check the page response and domain signals.
The first look starts now. The seven-day trial starts only when you ask me to keep watching.