# AI-powered monitoring Would you know if your checkout form stopped working overnight? Or if a recent deploy broke your login flow? Traditional monitoring can't catch these issues - it only tells you if your site is up, not if it actually works. AI-powered monitoring lets you describe what should work in plain English, and we'll test it like a real user would - clicking buttons, filling forms, checking content. No scripts to maintain, no complex setup. Every account gets 5 free AI checks to try it out. [ Start monitoring ](https://ohdear.app/register) 10 day free trial No credit card required ## Verify what matters with natural language prompts Skip the complex configuration. Just describe what you want to verify in plain English and our AI figures out how to check it. Check that forms work correctly, verify content displays as expected, ensure buttons link to the right pages, monitor API responses, or track DNS records. If you can describe it, our AI can check it. ## No code, no scripts, no headaches Just describe what you want to check Forget writing Selenium scripts or maintaining Playwright tests. A typical automated test takes 30-60 minutes to write and breaks every time your HTML changes. With AI-powered monitoring, you write what you want in plain English. Our AI figures out the selectors, handles the interactions, and adapts when your HTML changes. You just describe what should happen. Time savings example- Traditional script: 50 lines of code, 45 minutes to write, breaks when HTML changes - AI check: 2 sentences, 2 minutes to write, adapts automatically - If you're manually testing after deploys or maintaining test scripts, AI-powered monitoring can save hours per week ## Replace manual testing without writing code The middle ground between uptime monitoring and test suites Most monitoring tools only check if your site returns a 200 status code. That's not enough - your site can be "up" while the checkout is broken. End-to-end test frameworks like Selenium work, but they're a pain to maintain. Every time you change a button ID or restructure your HTML, tests break. You need developers to fix them. AI-powered monitoring sits right in between - it tests like an end-to-end suite but adapts to changes like a human would. No brittle selectors, no test maintenance, just plain English descriptions of what should work. ## Control your AI costs with smart scheduling Run expensive checks less often, critical checks more frequently AI checks cost money - whether you use your own provider or ours. But not everything needs to run every minute. Your checkout flow? Check it every 5 minutes - it's business-critical. Your about page content? Once a day is plenty. A DNS record you set once? Weekly is fine. Set your AI checks to run anywhere from every minute to once every 60 days. Every check result shows exactly how much it cost in tokens, so you know what you're spending. No surprises. ## AI that clicks and types Catch issues traditional monitoring misses Our AI controls a real browser and interacts with your website like a real person would. This is the only way to catch issues traditional monitoring misses: - Your login form returns a 200 status but doesn't actually log you in - JavaScript errors prevent the checkout button from working - Your search returns results but they're all wrong - A third-party script broke your navigation menu Our AI clicks buttons, fills in forms, navigates between pages, and checks what JavaScript logs to the console. If something would frustrate a real user, our AI will catch it. Our AI can also check DNS records and knows the current date and time for time-sensitive checks. ## Our AI remembers what it saw last time Track changes over time, not just current state Every time our AI runs a check, we give it the results from the previous run. This means it can compare what it sees now with what it saw before - opening up powerful use cases that go beyond simple "is it working" checks. Monitor for new RSS items appearing in your feed. Get notified when a product price drops by more than 25%. Track when new blog posts are published. Alert when your competitor changes their pricing page. Our AI has memory, so you can ask it to watch for changes over time. Traditional monitoring tools can only tell you about the current state - ours can tell you what changed since last time. Example memory-based checks- Notify me when a new RSS item appears at https://example.com/feed - Alert me when the price of this product drops more than 25% - Tell me when a new blog post is published to the homepage - Watch for changes to competitor pricing on their website - Track when new job listings appear on a careers page ## Screenshots show exactly what broke See your site through our AI's eyes When something goes wrong, you get a screenshot showing exactly what our AI saw at the moment it failed. Not a generic error page - the actual state of your site when the check failed. See the form that wouldn't submit, the missing product image, the error modal that appeared, or the broken layout. Debug in seconds instead of trying to reproduce the issue yourself. Our AI captures these automatically whenever it detects a problem, so you always have visual proof of what went wrong. ## Get notified when our AI detects an issue Notifications through any supported channel Our AI doesn't just detect issues - it decides when to notify you based on your prompt. Tell it to alert you only for critical problems, or ask for a notification whenever something specific changes. You're in control of what triggers an alert. When a notification goes out, it arrives through whatever channel works best for your team - Slack, email, SMS, Discord, PagerDuty, or any of our other supported integrations. You'll hear from us when something breaks and again when it's fixed. No noise, just the alerts that matter. ## See what our AI is thinking Detailed reports for every check run Every time our AI runs a check, you get a detailed report showing exactly what happened. See the full AI response, which tools it used, how many tokens were consumed, and whether it decided to send a notification or not. This transparency helps you understand how our AI interpreted your prompt and what it found. You're never left guessing why a check passed, failed, or triggered an alert. ## Who should use AI-powered monitoring? Perfect for e-commerce, SaaS, agencies, and content sites **E-commerce sites:** Monitor checkout flows, product displays, search functionality - all the things that directly affect revenue. **SaaS applications:** Verify signup flows work, dashboards load correctly, and critical user actions complete successfully. **Agencies:** Set up monitoring for client sites without writing custom code for each one. Just describe what should work in plain English. **Content sites:** Check that articles display correctly, comments work, and newsletter signups function. **Anyone doing manual testing after deploys:** Stop clicking through your site after every release. Let our AI do it for you. ## What AI-powered monitoring can and can't do It's powerful, not magic AI-powered monitoring is excellent for checking user-facing functionality, but it has limits. Here's what to expect: Great for- Testing forms and user flows - Verifying page content - Checking API responses - Monitoring configuration changes - Catching JavaScript errors Not ideal for- High-frequency checks (costs add up quickly) - Testing behind feature flags you control - Replacing comprehensive QA testing Think of it as continuous smoke testing - catching the obvious breaks fast, not replacing your full test suite. ## Real world prompts Monitor ## 1. Describe what you want to monitor The more specific you are, the better the results. ## Stop worrying, start monitoring Start a no-strings-attached 10-day free trial. You're all set in less than a minute. (No credit card needed.) Not convinced yet? Need help? Get in touch via . ### How AI-powered monitoring works We give you free checks to try the feature, then you choose how you want to pay - use your own AI provider or ours. Here's everything you need to know about getting started and how the system handles your checks. ## Start for free 5 AI checks on us **Every Oh Dear account includes 5 free AI checks to get started.** That's 5 complete test runs - enough to set up a few monitors and see how they work before choosing how you want to continue. Try monitoring your contact form, testing your checkout flow, or verifying your homepage content. No setup required, no credit card needed. Once you've used your free checks, choose to either connect your own AI provider or use Oh Dear's - whatever's most convenient. ## Choose how you want to pay for AI Use your own provider or let us handle it Once you've used your free checks, you have two options: **Bring your own AI provider:** Connect your own OpenAI, Anthropic (Claude), or Google Gemini account. You pay your provider directly based on actual usage. Just add your API key to Oh Dear and you're ready to go. **Use Oh Dear's AI providers:** Don't want to manage API keys? We'll handle it for you. We charge per usage on your Oh Dear bill - simple and convenient. Whatever works best for you. Switch between options anytime. ## You control the costs Only pay for what you use **AI checks cost money based on how many you run and how complex they are.** A simple check might cost fractions of a cent, while a complex browser automation could cost a few cents. If you use your own AI provider, you pay them directly. If you use Oh Dear's providers, it appears on your Oh Dear bill. Either way, you only pay for actual usage. Every check result shows exactly how many tokens were used, so you always know what each check costs. Set check frequencies that make sense for your budget - run critical checks every minute and less important ones once a day. ### From 50 lines of code to one paragraph See what you no longer need to write Here's what testing a contact form looks like with traditional Selenium automation: ``` from selenium import webdriver from selenium.webdriver.common.by import By from selenium.webdriver.support.ui import WebDriverWait from selenium.webdriver.support import expected_conditions as EC driver = webdriver.Chrome() driver.get("https://example.com/contact") # Wait for and fill the name field name_field = WebDriverWait(driver, 10).until( EC.presence_of_element_located((By.ID, "name")) ) name_field.send_keys("Test User") # Find and fill email field email_field = driver.find_element(By.CSS_SELECTOR, "input[type='email']") email_field.send_keys("test@example.com") # Fill message textarea message_field = driver.find_element(By.NAME, "message") message_field.send_keys("This is a test message") # Click submit button submit_btn = driver.find_element(By.XPATH, "//button[@type='submit']") submit_btn.click() # Wait for success message success_msg = WebDriverWait(driver, 10).until( EC.visibility_of_element_located((By.CLASS_NAME, "success-message")) ) assert "Thank you" in success_msg.text, "Success message not found" driver.quit() ``` **With Oh Dear's AI-powered monitoring, you write this instead:** Visit https://example.com/contact and fill in the form with name 'Test User', email 'test@example.com', and message 'This is a test message'. Click the submit button and verify a success message appears saying 'Thank you'. Alert me if the form doesn't submit or the success message doesn't appear." placeholder="Enter the prompt for the AI to check your site... No selectors to maintain, no waits to configure, no brittle XPath expressions. Just describe what you want to test in plain English. ## Really into the nitty gritty details? Want to get into the nitty gritty details of our monitoring? Have a look at our documentation. ## Frequently asked questions ### What is AI monitoring and what can it check? You describe what you want to verify in plain English, like "the pricing page loads and shows three plans" or "I can add a product to the cart and reach checkout." An AI then drives a real browser through your site, clicking buttons and filling forms, and tells you whether your expectation held up. ### How is this different from a normal uptime check? An uptime check tells you the page returned a 200. AI monitoring judges whether the page is actually right: the prices are correct, the form submits, the right content is there. It catches the broken things that still return a perfectly healthy status code. ### Which AI models can I use, and can I use my own account? You can bring your own OpenAI, Anthropic (Claude), or Google Gemini key, or use Oh Dear's. Every check shows exactly what it cost in tokens, so there are no surprises, and you get five free checks to try it first. ### How often do AI checks run, and what happens when one fails? Run them anywhere from every minute to once every 60 days. When a check fails, you get an alert with the AI's reasoning and a screenshot of what it saw, so you're not left guessing at what went wrong. [See all other FAQ items →](https://ohdear.app/docs/faq) ## Start monitoring instantly Start a no-strings-attached 10-day free trial. You're all set in less than a minute. (No credit card needed.) Not convinced yet? Need help? Get in touch via .