Changelog

What's new in BugOtter

New things, fixes and small wins, newest first. Have an idea? Tell us.

An otter reading a book by a lantern
  1. Dodo Payments, and Polar's sandbox

    Connect Dodo Payments with a read-only key and its payments, refunds, subscriptions and MRR show next to your other providers, without the tax Dodo collects. Polar sandbox tokens now work too, marked Test mode.

    Connect a payment provider →

    Groups: sort your apps by what they are to you

    Put apps in groups like Making money and Side bets, or one per client. Home shows a section per group with its totals, and a chip narrows Home, Analytics, Uptime and Errors to one group. Each group can be as loud as you like (everything, problems only, or outages only), an alert channel can be limited to a group, and agents and API keys can be limited to whole groups. Optional: skip it and nothing changes.

    How groups work →
  2. Each day on the traffic chart now says what its visitors were worth

    Hover a bar on Analytics: next to the day's revenue you now see how much was new money and how much was renewals, plus revenue per visitor and how many visitors became new customers. The last two show once a day has 30 visitors, so one sale doesn't swing them.

    Clearer churn, recovered Stripe payments, and customers in the API

    Churn is now named for what it measures: MRR churn, with the share of subscribers lost beside it, and net revenue retention under More numbers. Stripe payments that failed and then went through on a retry now count as paid later, instead of disappearing. The REST API adds the same ratios, failed payments paid later and the 30-day forecast to /revenue, and a new /revenue/customers lists customers one by one, with the provider's customer id and never a name or email.

    Customers in the API →

    Revenue you can trust: test data stays out, broken syncs speak up, and bad news gets an alert

    Once a live payment account is connected, test-mode accounts pause and their test charges leave every number. If a provider stops accepting your key, or syncing fails for 6 hours, you get one alert and a banner on Revenue. New alerts: MRR down 10% in a week, one big cancellation, and an app whose sales went quiet. Failed payments now alert at two in a week, not one expired card. Free plans no longer count as paying subscribers, and trials have their own tab in Customers.

    Revenue alerts →
  3. See who a customer is, in your own payment dashboard

    Open any customer on Revenue → Customers and click Open in Stripe to land on that customer, or Find in Paddle, Polar or Creem with their customer id ready to paste. BugOtter still keeps no names or emails: they stay with your provider.

    Lemon Squeezy can no longer be connected

    Lemon Squeezy is closing, so it's gone from the list of payment providers. If you connected it before, it keeps syncing while its API still answers. When you move to Stripe, Paddle, Polar or Creem, link the new product to the same app and its revenue history carries on.

    Webhooks: hear about deploys, outages and new errors the moment they happen

    Add a URL in Settings → API & webhooks and BugOtter posts signed JSON to it when a deploy ships, an app goes down or comes back, or a new error shows up (or one you fixed returns). Pick the events and the apps. Each request is signed so you can check it came from BugOtter, and a failed one is retried for about 9 hours. Send a test, and see what your server answered for the last 15.

    Set up a webhook →

    The API now shows every page's whole story, and every app at once

    New endpoints give each page's visitors, signups, revenue, real-visitor speed (LCP, INP, CLS), 404s, visits from AI assistants and which crawlers read it, plus what ChatGPT, Claude and Perplexity read and send you, and which pages answer "not found". Top lists can add signups, conversion and revenue per row. Leave out the app and stats, uptime and deploys answer for all your apps together. Keys moved to their own place, Settings → API & webhooks.

    What's in the API →

    Choose which apps a connection can see, right when you allow it

    When an agent or another app asks to read your BugOtter, the sign-in screen now lets you pick all apps or only some before you click Allow. Apps can also offer "Connect BugOtter" now: a standard sign-in instead of a pasted key, and you can disconnect them any time.

    Sign in with BugOtter →
  4. Alerts when your app gets slow, and reminders when it's still down

    Slow is now measured against each app's own usual speed: 5 minutes at 5× its normal response time sends a warning that says how slow, against what, and whether the time goes to your server, DNS, TLS or one region; another tells you when it's fast again. Prefer a fixed number? Pick 1 to 8 seconds in the app's Settings. An outage that's still open after 30 minutes sends a reminder, again at 2 hours, then daily. Down alerts now name the URL, what it answered and the start of the page it served, with a Fix with agent button.

    About slow alerts →

    The API now covers uptime, errors, revenue and deploys, as JSON or CSV

    The Stats API grew into the API: new endpoints for every app's health at a glance, uptime per day with incidents, MRR and revenue with new and churned MRR, open errors and deploys. Add format=csv to any of them for a spreadsheet-ready table; a short Google Sheets script turns them into a =BUGOTTER() formula. Same keys as your agent, read-only, on every plan.

    Put your numbers in a dashboard →

    Browser errors from your signed-in pages, and readable stacks on Next.js 16

    Add data-errors="everywhere" to the script tag to catch errors on pages you keep out of analytics, like your app's dashboard, without counting visits there; data-errors-exclude keeps pages with a secret in the URL out. Source maps now match Next.js 16 builds on Vercel, which serve files from /_next/static/immutable/. Database query errors keep the query but drop the values sent with it, and minified error names read as Error, so one error stays one group across deploys.

    About error tracking →
  5. Bring your history from DataFast, Fathom, Simple Analytics and Google Analytics

    Bring your history now reads straight from DataFast, Fathom and Simple Analytics with an API key that's used once and never stored, and from Google Analytics with a one-time, read-only Google sign-in. It goes back month by month until your history runs out. Plausible and Umami exports work as before.

    About importing history →

    How people buy, and customers who stay

    Two new views under Go deeper in Analytics, from payments already linked to visits. How people buy shows the usual path to paying and the pages buyers read far more than anyone else. Customers who stay groups subscribers by the month they started and shows who still pays months later, by where they came from.

    About journeys and cohorts →

    See the bots, and which AI reads your site

    Analytics has a Bots view under Go deeper. It lines up how often ChatGPT, Claude and Perplexity read your site with how many visitors they send, checks your robots.txt and llms.txt every day, and shows the bots kept out of your numbers: which ones, from which network, on which pages, with a warning when they hammer a login page. A few lines on your server add crawlers that never run scripts, like GPTBot and Googlebot, and the 404s they hit. And the first visitor an AI assistant sends you is now a milestone you can share.

    About bots and crawlers →
  6. Build funnels from your real pages

    New funnels start from ones suggested by your own traffic, like home to pricing to signup. Each step is a page or a goal picked from your real paths, and its visitors show as you type, so a typo shows up as 0 before you save.

    About funnels →

    Milestone cards worth posting

    Passing $1,000 MRR, or 10,000 visitors in a month, now comes with a card to share, alongside first sales and best weeks. Home offers it once when it happens, and you can hide the amount.

    About good-news alerts →

    Analytics at a glance

    Sources, pages, countries and devices now sit side by side, each with its revenue, paid rate and errors per row. Revenue is drawn as a line on the chart, next to the deploy and error marks. With only a handful of visitors, the page keeps to plain counts until rates mean something.

    Connect several payment accounts

    Connect more than one Stripe, Paddle, Polar, Lemon Squeezy or Creem account: one Creem store per app, or a co-founder's Stripe for a product you share. Name each one and pick the app it sells, and its products go to that app on their own. Each account syncs and disconnects separately.

    Add another account →
  7. Your agent can rename an app or follow a domain move

    With write access, your agent can rename an app or point it at a new URL, keeping its history. If adding an app would go over your plan, it tells you right away. When something can only be done in the app, it gives you the exact page. The sign-in screen now shows an agent's name and logo only when the sign-in goes back to that agent's own website.

    What your agent may do →

    Connect your agent in three steps

    Pick your agent, add BugOtter, sign in. Cursor and VS Code install with one click, Grok Bot adds it from a single message, and any agent can install it from a copied prompt. For an AI chat that can't connect, paste a snapshot of what needs you instead.

    Connect your agent →

    Paddle joins Stripe, Polar, Lemon Squeezy and Creem

    Connect Paddle with a read-only key to see revenue, MRR, refunds and failed payments next to your traffic. Pass the visitor id as custom data on the checkout to see which source paid. Creem one-time sales now show the product they were for.

    Connect Paddle →
  8. Autopilot checks Claude's fix in production

    Home now shows when Claude's pull request is ready to review. After you merge and deploy, BugOtter watches the app and tells you, and the GitHub issue, whether the fix held. If it didn't, one click asks Claude to look again. The weekly digest counts the fixes that worked.

    How the fix check works →
  9. Autopilot: Claude fixes it while you sleep

    Connect GitHub and pick each app's repo. When an outage or a broken flow doesn't go away, BugOtter opens an issue with what it measured and asks Claude to fix it. You wake up to a pull request. New errors open the issue too, and you start Claude with one click on Ask Claude.

    How Autopilot works →

    What it costs you, in dollars

    With revenue connected, Home, alerts and the weekly digest put a price on what needs you: an outage or broken checkout shows about how much was missed so far, and an expiring certificate or domain shows what a day without it would cost.

    About alerts →

    BugOtter's own status page

    BugOtter watches itself like any customer's app, every minute. See it at /status. If check rounds ever stop, the page says so.

    BugOtter status →

    One error inbox, and a dark website

    Errors live in one inbox instead of two, and Home leads with what is down. On a phone, Fix with agent can email the prompt to you for later. The website, docs and guides follow your system's dark setting.

  10. Help is one click away

    The new ? button in the top bar opens the docs, keyboard shortcuts, this page and support. Press ? anywhere in the app to see every shortcut.

    Contact support →

    Flow checks: a real browser walks your signup and checkout

    Show BugOtter the steps once, and a real browser clicks through them on a schedule. If signup or checkout breaks, you hear about it before your customers do.

    How flow checks work →

    Any coding agent can connect, Devin included

    The setup prompt now works with any agent that speaks MCP, with exact steps for Devin, Zed, Gemini CLI and more. Agents can also edit your flow checks.

    Connect your agent →

    A smoother first few minutes

    Setting up is three short steps, and a new step connects your coding agent right away, so it can add the tracking code for you.

    Sturdier alerts and revenue numbers

    Connecting Telegram is now one tap on a Start link. Stripe and Lemon Squeezy numbers match your own dashboards more closely, and alerts only go to channels you proved are yours.

    Alerts in the docs →

    Readable Next.js stack traces

    Upload source maps from a Next.js build and errors point at your real code. Sentry SDKs that compress their reports now arrive too.

    Upload source maps →

    Widgets and maker pages

    Show live uptime, traffic or revenue on your own site with a small widget, and list every app you've shipped on one public maker page.

    Widgets and badges →
  11. A clearer MRR chart

    The revenue waterfall shows new, upgrades, downgrades and churn as plain steps, so you see at a glance why MRR moved.

    Creem joins Stripe, Polar and Lemon Squeezy

    Connect Creem read-only and its MRR, customers and profit show up next to everything else.

    Revenue in the docs →
  12. Status pages on your own domain

    Put any app's status page on a domain like status.yourapp.com. Checks can now send POST requests, and error rates are exact.

    Set up a domain →

    Analytics that explains itself

    Clean visitor numbers, the paths people take, and imports from your old tool. Don't need a module? Turn it off and it leaves the app.

    Analytics in the docs →

    Sign in with Google

    Use your Google account to sign in, and connect Search Console to see which searches bring people to each app.