Privacy Policy — outage.me

Operator: Smart Signals LLC, a Wyoming limited liability company. Contact: privacy@outage.me. Effective: 12 November 2026.

outage.me shows the live status of online services, using each provider's own public status page. We collect as little as we can. This policy lists everything we do collect.

Public site (outage.me and its regional pages, such as outage.me/br and outage.me/de)

• No analytics, no ad trackers, no tracking cookies. We don't record your IP address anywhere. Apart from the request log described under Security and server logs and the device type in usage counts below, we don't record your browser details.
• Usage counts. When you view a page, click a service card or open search, we record which page and card it was (for example "github card clicked"), a timestamp, the referring website's domain, whether the device is a phone, a tablet or a computer, and whether the page was opened as an installed app. Your browser works out that device type and sends only that one word, never its own identification. We don't record who you are, your IP address or any text you type. These records are deleted after 30 days.
• Search box. If you type something that looks like a web address, it's sent to our server to check whether that service has an official status page we could track. It isn't linked to you.
• Speed test (optional). The test runs against Cloudflare (speed.cloudflare.com). We only receive results if you tick the consent box. Then we store your download and upload speed, latency, jitter, and your internet provider's network number and name. No IP address or location. Results are deleted after 30 days and appear only as per-provider averages.
• User reports (optional, where a service page offers them). If you click a problem on a service's page (for example "Mobile data"), we count it: the service, the problem type, a 15-minute time slot, your approximate location (country, state and city, as estimated by our content-delivery network from your connection; no GPS and no browser location request) and your internet provider's network number. We don't store your IP address or browser details with any report. To count each person once per service every 30 minutes, we keep a one-way code derived from your IP address and browser; its key changes every day, and the code is deleted after 30 minutes. Report counts are deleted after 35 days. We publish only totals: a place is shown only once it has at least 5 reports, and network numbers are never shown. Legal basis: our legitimate interest in showing whether a service is failing for other people too (GDPR Art. 6(1)(f); LGPD Art. 7, IX).
• Email alerts (optional). We store your email address and the services you picked, and send alerts from alerts@outage.me through Amazon SES. Nothing is sent until you click the link in our confirmation email; if you never do, your address is deleted after 7 days. Every alert has a one-click unsubscribe link, and unsubscribing deletes your address. We also keep a one-way hash of your email for 24 hours, to stop repeat sign-ups. Subscriptions made before 3 October 2026 were moved from Amazon SNS to this list.
• Stored in your browser, never sent to us: the services you pinned, the cards you dismissed, and your last-used tab.
• Saved on your device for offline use, never sent to us: where your browser supports it, outage.me keeps its icons and a copy of the last pages you opened, so a page still opens without a connection (it then says you're offline and when it was last updated). On app.outage.me the saved pages are deleted when you sign in or out, and are never shown once they are 12 hours old. Clearing this site's data in your browser removes them.
• Our own staff marker. Opening a page with ?internal=1 sets a cookie, ot_int=1, that only tells our counters to leave that browser's visits out (we use it so our own testing is not counted as visitors). It is the same value in every browser that has it and identifies no one. ?internal=0 removes it. We never set it otherwise.

Pro app (app.outage.me)

• Account. Sign-in uses Amazon Cognito, with email, Google or Apple. We keep your email address. From Google or Apple we receive only your email and whether it's verified; if you choose Apple's "Hide My Email", we get the relay address.
• Your settings. Your watchlist (providers, regions, network numbers, vendors, alert threshold) and any alert channels you add (email, a Slack or webhook URL, a PagerDuty routing key). We use these only to send the alerts you asked for, to the destinations you configured. An email channel gets the same confirmation email and one-click unsubscribe as email alerts above, and removing the channel deletes the address from our alert list.
• Ask. Your questions go to Amazon Bedrock (AWS) to generate answers. Each question and answer is stored for 1 hour, only so follow-up questions work, then deleted automatically. A daily question counter is kept for 2 days.
• API keys. We store only a one-way hash of each key, its label, its last 4 characters, and when it was created and last used.
• Usage counts. Page views, tab switches and settings changes are recorded with your account ID and the same device type (phone, tablet or computer; installed app or not), for 30 days. Your question text is never recorded.
• Cookies. Sign-in cookies (strictly necessary) last 12 hours: CloudFront-Policy, CloudFront-Signature, CloudFront-Key-Pair-Id, tarantula_principal and tarantula_tier. A session cookie (tarantula_rt, strictly necessary) keeps you signed in on your device: it holds a sign-in renewal token from Amazon Cognito, can't be read by page scripts, and is sent only to our sign-in service, which uses it to renew the 12-hour cookies. It expires after 14 days without use and at the latest 30 days after you signed in; signing out deletes it and cancels the token. A 10-minute cookie (tarantula_oauth_state) protects the sign-in step, a 10-minute cookie (tarantula_next) remembers which page to open after you sign in, and a 10-minute cookie (__Host-oauth_csrf) protects the "Allow" button when you connect an app. A 10-minute cookie (tarantula_reauth) is set when you have just signed in; deleting your account needs it.
• Connected apps (MCP). If you connect an AI app (for example Claude or ChatGPT) to your account, we store which app you approved, when, and when it last used access. The app receives tokens that are stored only as one-way hashes on our side. They expire (access after 1 hour; renewal after 30 days, or after 7 days unused). Disconnect any app, or all of them, under Connected apps in the Pro app. Questions an app sends through the answer tool are handled exactly like Ask above.
• Contributing outage observations (optional, off by default). If your organization turns on Contribute outage observations, your software (an AI app, a hook, or an OpenTelemetry collector) can tell us that calls to a public service failed. We store your organization's account ID, the public hostname (only hosts on our published list), the type of failure (for example a timeout or a 5xx error), a 15-minute time slot, call and failure counts, and, if your software sends it, the cloud region it runs in. We never receive or store URLs, paths, request or response content, error messages, or your own internal hostnames. We use this to confirm outages across organizations, to credit your organization with extra questions when a report is confirmed, and to keep a per-organization trust score. Records are deleted after 35 days at most. We publish only totals across at least 3 organizations, never which organizations reported, and a region only once at least 5 organizations reported from it. Turn it off at any time on the Keys page. Legal basis: our legitimate interest in detecting outages, and the contribution you choose to make (GDPR Art. 6(1)(f); LGPD Art. 7, IX).

Push alerts

• On a service page ("Get push alerts"), no account needed. Your browser asks for permission only after you tap. We then store what your browser's push service gives us for this site: a push address and two encryption keys. With them we store the services you follow on that page's site (at most 20), which version of the site it is (for example outage.me or outage.me/br), and the day the subscription was created and last changed. We don't store your IP address, your browser details or anything else with it. Your browser sends the push address in the body of the request, never in a web address, so it never appears in our request logs. Your browser also keeps its own copy of the services you follow, to show the button's state.
• On app.outage.me ("This device" alert channel, Pro). The same push address and keys, linked to your account and to the exposure profile you added the device to, the kind of device your browser works out (such as "iPhone"), and the day you added it. The device list in the Pro app shows that label and that day.
• In the outage.me app (iPhone, Android), no account needed. When you tap a star or "Get push alerts", the app asks your phone for permission. We then store the app's push token, which Expo (the app's push provider) issues for your installation of the app, a random code the page creates to prove later requests come from the same app, the services you follow on that region's board (at most 20), which region it is, and the day it was created and last changed. We don't store your IP address, your device details or anything else with it.
• What a notification says. Only a service's name and its state, for example "GitHub: degraded", and a link to that service's status page. A Pro notification never names your profile, your watchlist, or any region, network or vendor of yours, because phones show notifications on the lock screen.
• Delivery. The push service run by your browser's maker delivers each notification: Google (Chrome, Edge and Android), Apple (Safari, and outage.me added to an iPhone or iPad Home Screen), Mozilla (Firefox) or Microsoft. Each notification is encrypted so that only your browser can read it; the push service sees only that a message was sent to that push address, when, and its size. In the outage.me app, notifications go through Expo (650 Industries, Inc.) to Apple or Google, which deliver them to your phone; Expo and they see the notification's text (a service's name and its state) and the app's push token.
• Stopping. Tap "Stop push alerts" on the page (unfollowing your last service deletes the subscription), remove the device in the Pro app, delete the exposure profile, or turn off notifications for the site in your browser or phone settings. In the app, un-star the service or turn off notifications for outage.me in your phone's settings. When a push service tells us a subscription no longer exists (for example after you uninstall the app), we delete it. Deleting your Pro account deletes your devices.
• Legal basis: your consent, given when you allow notifications (GDPR Art. 6(1)(a); LGPD Art. 7, I). You can withdraw it at any time, as above.

The outage.me app (iPhone, Android)

• Error and stability counts. So we can find a broken release, the app tells us when a page inside it hits an error, when one of our requests fails, and when the app starts or a previous run ended unexpectedly (for example a crash). Each report contains only that kind of event, the app's version and whether it's the iPhone or Android app; for an error, the error's type (such as "TypeError"), never its message; for a failed request, the kind of request (such as /api/push, with numbers and codes removed) and its response code. We don't receive your IP address, your account, your device details, anything on the page or anything you type. We keep these reports in our server logs for 30 days and use only their totals. To know whether the previous run ended unexpectedly, the app keeps one small file on your phone with its version and whether it was last closed normally; deleting the app deletes it. Legal basis: our legitimate interest in keeping the app working (GDPR Art. 6(1)(f); LGPD Art. 7, IX).

• Updates. When it starts, the app asks Expo (650 Industries, Inc.) whether a newer version of its own code is available, and downloads it if there is one. The request carries the app's version, its update channel and whether it's the iPhone or Android app. Like any server, Expo sees your IP address. We receive nothing from these requests.

Browser extension (outage.me for Chrome and Edge)

• What it downloads. Every 2 minutes the extension downloads two public files from outage.me: the list of services that currently have a problem, and the list of website addresses that belong to each tracked service. It compares the site you are on with that list inside your browser, to colour its icon. It doesn't send us which sites you visit.
• The open tab, only when you click. When you click the extension's button, it reads the open tab's address and, on a page you ask it to check, how long that page took to load and which hosts it loaded from. That stays in your browser unless you press a button that sends something, and each such button says what it sends: "Is it just me?" sends the site's name (not the page) to outage.me, which then loads that site's home page from five locations (and, for a service whose videos or images come from other servers, such as X, one small public file from each of those servers); "Monitor this" sends a service's name; search sends what you typed only when it looks like a web address, as on the site.
• User reports. "I have a problem" sends the same report as on outage.me (see User reports above), with the same counting rules.
• Which network you are on (only when you check why a video or image isn't loading). To tell whether the problem is the service, its delivery network, your internet provider or your VPN, the extension asks outage.me which kind of network you are connected through. We answer from the network number our content-delivery network sees (a number that identifies a network, such as an internet provider, a cloud or a VPN service, never a person or a household): its kind and name, for example "your internet provider, Claro" or "a VPN network, Proton VPN". We keep nothing from this request. The extension remembers the last result for each network in your browser only, so it can compare, for example "through your provider last time: 14 Mb/s; through this VPN: 0.3 Mb/s".
• Following services. "Follow" works like push alerts on outage.me (see Push alerts above), with one difference: a notification from the extension also shows the provider's own title for the incident (for example "Incident with Actions"), because a computer's notification is not a phone's lock screen.
• Signing in (Pro, optional). If you sign in from the extension, it keeps the sign-in tokens in your browser's extension storage, never in a web page. Questions you ask go through Ask (see Pro app above), with the open site's name if you choose to include it.
• Monitor (if your account uses outage.me Monitor). While you are signed in, the extension asks outage.me every 2 minutes for your Monitor incidents, to show them on its icon and as notifications; acknowledging or adding a note from the extension is recorded in your account's audit log, as in the Pro app. "Monitor this page" sends the open page's address (without anything after a #) and, only on that click, a word from its title, to create a check in your account; you can edit both before saving. "Test now" runs that check once from our locations and stores nothing. "Attach what I see" loads that incident's check address once from your browser, without cookies, and adds the result (up or down, the HTTP status and timings, never the page's content) to the incident's timeline in your account, under your name. Chrome asks for permission for that one site, and the extension hands it back right after.
• Developer panel. In your browser's developer tools, the outage.me panel groups the open page's failing requests by provider. Request addresses stay in the developer tools; only "Report what I see" sends a report, as above.
• Helping others see outages (optional, off by default). Three separate switches in the extension, none of which does anything until you turn it on. The first two ask your browser for permission to see requests on every site, because that is how the extension notices that a service is failing; it ignores every site that isn't on outage.me's published list, at the moment it sees the request.
• Count errors on tracked services (optional). Every 15 minutes the extension sends the host name of a tracked service that failed in your browser (for example api.stripe.com), the kind of failure (DNS, timeout, TLS, connection, server error or other) and how many times. Never a page, a path, anything after the host name, a page title, or any site outside the list.
• Share page load times of tracked services (optional). When you open a tracked service's own website, the extension sends how long the page took, only as a range (for example "0.5 to 1 second"), never an exact time. Nothing runs inside the page: the time comes from your browser's own request timing.
• Attach the error I saw to my reports (optional). When you send a user report, it also carries one word for the kind of failure your browser saw for that service in the last 30 minutes.
• What we keep from these three. We match each host name to its service and then discard the host name. We keep only totals per service and per 15-minute slot. We don't store your IP address, your browser details or your location with any of it. When the extension reports errors, we also keep your network's number with the error totals (a network such as your internet provider or VPN service, never a person), so a failure concentrated on one network shows up, for example "12 browsers on one provider see this, none elsewhere". A network is shown only once at least 5 browsers on it reported errors in the same slot, and then by its name and kind (internet provider, cloud or VPN). To count each browser once per service and slot, we keep a one-way code made from your IP address and browser; its key changes every day, and the code is deleted after 24 hours. We publish a slot only once at least 5 browsers contributed to it, and never which browsers. Totals are deleted after 35 days.
• What you can see, and stopping. The extension shows when it last sent something and what it was (how many services, errors and page loads), and what is waiting to be sent. Turn a switch off to stop it; turning off both of the first two also removes the every-site permission. Removing the extension deletes everything it stored in your browser.
• Legal basis for these three: your consent, given by turning a switch on (GDPR Art. 6(1)(a); LGPD Art. 7, I). You can withdraw it at any time by turning the switch off.
• Stored in your browser, never sent to us: your region, the services you follow (the push service also has the list, see above), quiet hours, and the counts that are waiting to be sent.

Third parties

• Amazon Web Services: hosting, sign-in, email delivery and the AI model.
• Google, Apple, Mozilla or Microsoft push services: they deliver push alerts, only if you turn them on.
• Expo (650 Industries, Inc.): relays push alerts to the outage.me app, only if you turn them on in the app, and delivers the app's updates.
• Google and Apple: only if you sign in with them.
• Cloudflare: the speed test, only if you run it.
• Google (Chrome) or Microsoft (Edge): if you install the extension from their store, their store and their push services, as above.
• Google Fonts: the pages load fonts from Google, which receives your IP address the way any web request does.
• Slack, PagerDuty or your own webhook: only if you configure them as alert channels.
• We don't sell or share personal data for advertising.

Security and server logs

Our edge firewall rate-limits requests by IP address, and AWS keeps a small sample of blocked requests. Our servers log errors, but not IP addresses, email addresses or anything you type. Server logs are deleted after 30 days.

Our content-delivery network keeps a request log for outage.me and app.outage.me: the time, the page address (without anything after a ?), the response code and size, how long it took, the referring page, your browser's user-agent string, and the country and network number estimated from your connection. It never contains your IP address or cookies. We use it only to see overall traffic, errors and automated visitors, and it is deleted after 30 days.

Your rights

You can delete your account yourself: in the Pro app or the outage.me app, open Account and choose Delete account. You sign in again to confirm, and the deletion happens at once. Your sign-in, your settings and alert channels, API keys, webhooks, connected apps, devices, and the usage counts recorded with your account ID are deleted. Ask questions are already deleted after 1 hour, and our database backups roll over within 35 days. If you have an active paid plan, cancel it first. You can also ask for access to your data, a correction, or deletion of your account and data by writing to privacy@outage.me. We answer within 30 days. If you are in the EU, the UK, Brazil, India or Singapore, you also have the rights your local law gives you (GDPR, UK GDPR, LGPD, India's DPDP Act, Singapore's PDPA), including the right to complain to your data-protection authority; send those requests to the same address.

Leia esta política em português.

Diese Erklärung auf Deutsch lesen.