Tenty watches uptime, deploys, payments, AI spend, database limits and security, spots what broke and suggests the fix, on a schedule and after every deploy. You keep shipping, someone has to stay up
Next to fix: prune old uploads or raise the storage limit
Example board, made-up values
What Tenty monitors
Five watches, one light and a suggested fix for whatever turns yellow or red. Each reads your app from outside or through a read-only key you hand over
Up and answering
Tenty notices before your users open a ticket
HTTP checks every minute on paid plans, with status, header, JSON field and latency assertions. When the app goes down, Tenty reads your providers' status and tells you whether the outage is yours or theirs. A public status page shows uptime and incidents to whoever asks
Uptime
HTTP assertions
Outage source
Status page
Every deploy, walked
Shipped on Friday? Each deploy gets a real browser, a real inbox and a real checkout
Your pipeline calls the Tenty hook when a deploy finishes. Tenty loads your key pages in a headless browser and checks for blank screens, page errors and must-have text, signs up with a real confirmation email, walks checkout through to a paid subscription and diffs the deploy against the last one: scripts, outside hosts, pages and DNS records. Screenshots and the raw log stay with each run
Tenty hooks
Scan diff
Key pages
Signup and login
Checkout walk
DNS changes
Money in, money out
A payment is a promise your app has to keep
A restricted, read-only Stripe key lets Tenty watch renewals, failed charges, cancellations and webhook delivery, and flag when payments go quiet. A read-only key from your AI provider lets it track AI spend, alert on a spike and tie it to any AI route the security check found open
Subscriptions
Webhook delivery
AI bills
Database watch
Limits arrive on schedule, backups prove themselves on a bad day
Through your Supabase connection, Tenty tracks database size, storage, bandwidth and monthly active users against your plan and projects the day each limit lands. It also checks that backups exist and restore, so the bad day stays a short one
Usage limits
Restore check
Security, from outside
Every security check is free, forever
Deterministic checks read your pages, bundles and DNS the way a stranger would: live secret keys in the frontend, AI endpoints anyone can call, open Supabase and Firebase data, source maps, headers, cookies and CORS, vulnerable libraries with CVE ids, subdomain takeover and email records. Each finding carries the raw request, a curl that reproduces it and a fix prompt
Secret keys
Open AI endpoints
Supabase and Firebase
CVE matching
Subdomain takeover
SPF, DKIM, DMARC
How it works
1
Verify the domain
A DNS TXT record or a file on your server. Results reach the verified owner only
2
Connect what you want watched
The hook URL in your CI, a restricted Stripe key, an AI provider key, a Supabase connection. Each is optional, read-only and revocable, and the chat in the app walks you through each setup
3
Read one light, fix what it shows
Red, yellow, green or grey, and the next thing to fix. Every finding and every report, from any watch, carries a suggested fix, and the chat answers what the fix leaves open. Alerts go to email, Slack, Discord, Telegram, web push or your own webhook
High
Signup stopped working with the 14:02 deploy
Tenty tried
Sign up with a fresh inbox
Your app answered
500 on POST /auth/signup, no confirmation email after 10 minutes
POST /auth/signup 500body: {"error":"relation \"profiles\" does not exist"}screenshot: signup-1402.pngprevious deploy, 11:40: signup passed
Expected a confirmation email within a minute
Reproduce it
curl -i -X POST https://yourapp.com/auth/signup -d 'email=you@example.com'
Suggested fix: run the profiles migration on production
You
Why did the migration skip production?
Tenty
Most likely your production build runs without the migrate step: the 11:40 deploy had no schema change, the 14:02 one added profiles. Add the migrate command to the production build, redeploy, and the next walk closes this finding
The chat answers from your app's results
Fits the pipeline you already have
One hook URL after each deploy, alerts where your team already reads and a token for your own scripts. Recipes for GitHub Actions, GitLab CI, Vercel, Netlify and Cloudflare wait in the app
Examples from every watch, each with the proof, what it means and a suggested fix. Examples use made-up values, your real ones stay yours
Medium
Outage source, 03:12
Your app was down for 4 minutes
Your provider had a database incident, your app came back with it
03:12 yourapp.com timed out
03:12 provider status: degraded database
03:16 back up, answered in 210 ms
The incident and its source are on your status page
Medium
Uptime check, latency budget 800 ms
Checkout answers slowly
Slow checkouts lose sales before anything breaks
/api/checkout answered in 2.4 s
Add an index or cache the query behind this route
High
Key pages, deploy at 14:02
The pricing page renders blank
Visitors land on an empty page and leave politely
GET /pricing 200, 0 visible words
TypeError: plans.map is not a function
must-have text "Start free" missing
Fix the plans fetch, the screenshot is attached
High
Subscriptions, checked hourly
Your Stripe webhook is falling behind
Customers pay and your app may never hear about it
evt_1QxA still waiting for delivery after 41 minutes
last attempt: 500 from /api/stripe/webhook
Check the signing secret in your environment
Medium
Subscriptions, checked hourly
Renewals went quiet
Six renewals by noon on a usual day, none today
renewals today: 0, usual by noon: 6
last invoice.paid 19 hours ago
Check the webhook endpoint and your Stripe dashboard
High
AI bills, linked to a security finding
AI spend tripled since noon
Someone found an open route and is enjoying your credits
spend since 12:00: $38.40, usual: $11
route open to anyone: POST /api/chat
Put the route behind sign-in and rotate the key
Medium
Usage limits, read daily
Storage reaches the plan limit in 9 days
Uploads stop working the day it fills up
storage: 8.1 GB of 10 GB
growth: 0.2 GB a day
Prune old uploads or move to the next plan
Critical
Owner-verified test, as a signed out stranger
Anyone can read your profiles table
Emails and billing details are one request away
GET /rest/v1/profiles?select=*
HTTP/2 200, content-range: 0-2313/2314
Turn on row level security and add a policy
Low
Security headers
Two headers are missing
Browsers get fewer instructions on how to protect your pages
Content-Security-PolicyMissing
Strict-Transport-SecuritySet
X-Content-Type-OptionsSet
Referrer-PolicyMissing
Copy the header block into your host config
Your week, in one email on Monday
Still open: storage reaches its limit in 9 days. Suggested fix: prune uploads older than 90 days
The week at a glance, with a suggested fix for everything still open
99.98%
Uptime
9
Deploys walked
112
Renewals
$71
AI spend
2
Findings closed
Security checks, free forever
Every security check is free on every plan for the verified owner. Try the first look on any address, the full result waits behind ownership
Stripe test secret keys
Leaked API keys
Supabase service_role keys
AI keys and calls in the browser
Open database tables
API operations that answer without a login
Public storage buckets
Anonymous sign-in on Supabase
Exposed source maps
Exposed .env and .git files
Missing security headers
CORS mistakes
SSL and domain expiry
The first look is free and takes about 20 seconds
Questions
What does Tenty watch?
Your production apps: whether they are up, what each deploy changed, whether signup and checkout still work, whether payments keep coming in, what your AI provider and your database are charging you, and what a stranger can reach. It all lands as one light per app and the next thing to fix
Do the checks use AI?
Detection is plain code and gives the same answer every run, so a finding that shows up twice is the same finding. AI tailors the suggested fix to each kind of finding and powers the chat that answers questions about your findings and your setup
Can I ask Tenty questions?
The chat in the app answers questions about your findings and walks you through setups, from a restricted Stripe key to the deploy hook. It reads your app's results, so the answers are about your app, and it admits when the evidence runs out
What does Tenty send to my app?
GET requests at a calm pace, one at a time, from published addresses with the user agent TentyScanner. Deeper tests run on a domain whose owner proved it, with a permission that lasts 15 minutes. The toolbox page has the details
Does Tenty need my code?
Tenty reads the running app from outside, the way a visitor would, plus the read-only keys you choose to hand over. Your repository keeps its secrets, it has plenty
Which keys does Tenty ask for?
Read-only ones, and only for the watches you turn on: a restricted Stripe key, a read-only key from your AI provider, a Supabase connection for usage and backups. You can revoke each one in a click
How does Tenty learn about a deploy?
Your pipeline or your host calls one Tenty URL when a deploy finishes: a GitHub Actions or GitLab CI step, or a Vercel, Netlify or Cloudflare notification. Free accounts can use the hook with the basic action, and paid plans add every other action
How is Tenty different from a status page or a one-time scan?
A status page tells you the app is up, a scan shows today. Tenty walks signup and checkout after each deploy, reads your payments and bills, and returns on a schedule. When Tenty shouts wolf, it brings the wolf: the evidence and a command that reproduces it. Check it yourself, the village has heard enough boys
Which editors and builders does it fit?
Any app on a public address, written by hand or with Lovable, Cursor, Bolt or Claude Code. Fix prompts are plain text you paste into your editor, and the curl runs in any terminal
Where does the money come from?
Every security check is free on every plan, for the owner of a verified app. Tenty earns from the watches around it: faster uptime checks, deploy walks, payment and bill watch, alerts in Slack, Discord and push, and longer logs
Is Tenty finished?
We are in beta and we are figuring things out. We would rather ship it, watch it work on real apps and fix what it gets wrong than polish it in private. When a finding is wrong, press the button on it and a person reads your report
How do I get started?
Free accounts are open. Create one, add your app, prove it is yours and turn on the watches you want. Curious first? Try the security check on your address, further down this page