Product

Changelog

What changed in BugMojo, newest first. The browser extension updates itself through your browser’s store, and the web app updates the moment a change ships. Questions about a change? Get in touch.

Extension 1.1.6: no "all your data on all websites" warning

  • Extension
  • Security

Installing BugMojo no longer asks for access to every website. Instant replay asks for your sites one at a time, and a capture anywhere else starts from your click.

  • Chrome no longer shows "Read and change all your data on all websites" when you install BugMojo.
  • Instant replay asks for access only to the sites you choose: the site you turn it on for, your local dev hosts, or your project domains in a single prompt. "Every site" is still there if you want it, and it is never the default.
  • On any other site, the toolbar icon, the keyboard shortcut or the right-click menu captures the current tab. Console and network capture start at that click.
  • Removing a site’s access stops recording on that site.
  • If you already gave BugMojo access to every site, it keeps working and you will not see a new prompt.
  • Background tabs on sites that are not recording no longer run any BugMojo script.

Read more: How instant replay works · Why BugMojo asks for permissions

Extension 1.1.5: select an area, crop, and hand the bug to your AI agent

  • Extension
  • AI agents
  • Security

Faster captures, a crop tool, steps to reproduce written for you, and one-click ways to pass a bug to Claude Code or Cursor.

  • Select area: capture part of the page from the side panel or the right-click menu. The selection overlay never appears in the screenshot or the replay.
  • Crop in the annotator at full resolution. Arrows, shapes and blur now land exactly where you drag them, and each tool has a one-key shortcut.
  • "Record from now" puts Stop and Discard buttons on the page. Stop opens the review with the whole session, not just the last 30 seconds.
  • Steps to reproduce are written for you from the recording, and you can edit them. Anything typed into a field is never included.
  • Copy a ready-made prompt for Claude Code or Cursor, copy the report as Markdown, or download the network log as a HAR file.
  • Links are copied to your clipboard as soon as you create or share a report.
  • The keyboard shortcut and the right-click menu now work when you are signed out (Quick Capture) or have no team yet.
  • When a capture fails, you see the reason and a Retry button.
  • Fields you already hide from FullStory, Hotjar, Sentry, PostHog, LogRocket, Amplitude, Clarity, Datadog or rrweb stay hidden in BugMojo too.
  • With "Every site" turned on, BugMojo never records webmail, password managers, payment pages, banks or health portals.

Extension 1.1.4: instant replay only on your sites, and complete replays

  • Extension
  • Replay
  • Fixes

The always-on replay buffer now runs only on the sites you work on, and a recording bug that dropped page changes is fixed.

  • Instant replay, the rolling buffer that lets you capture what just happened, now runs only on your sites: local dev hosts, your workspace’s project domains, and sites you turn on with one click. Recording on every site is opt-in.
  • Tabs you are not looking at do not record. When you come back to a tab, a fresh recording starts.
  • Canvas recording is off by default and starts only on pages that have a canvas.
  • Pages that change constantly pause recording and try again two minutes later, so one heavy page does not slow down your browser.
  • Fixed: most page changes were being dropped, so replays showed the first frame plus scrolling and mouse movement. Replays now show the whole session.

Read more: How instant replay works

Verified fixes show up in your tracker

  • Integrations

BugMojo works alongside the tracker your team already uses. When a fix is verified, the linked issue there is closed with the proof.

  • When a bug is marked verified fixed, BugMojo closes the linked GitHub, GitLab, Jira, Linear or Azure DevOps issue, adds a verified-fixed label and posts one comment with the proof. Linked pull requests are never touched.
  • When a bug is reopened, the linked issue is reopened too, with the reason.
  • Fixed: a verified fix used to reopen the linked GitHub issue instead of closing it.
  • Tracker sync now also runs for bulk updates, CI results, the REST API and changes made in the extension.
  • The issue page shows which tracker issues are linked (for example GH #123 or LIN-45), and each tracker card has an auto-push switch.
  • Slack no longer posts every status change twice.

Read more: Integrations overview

A replay viewer that lines up with the logs

  • Replay
  • AI agents
  • Fixes

Console and network entries now share the replay’s clock, and the player takes you straight to the first error.

  • Console and network entries use the same clock as the replay, so they line up with what you see on screen. Dates no longer show up as 1970 on the issue page.
  • The player opens about three seconds before the first error. "Jump to first error" and red markers on the seek bar take you straight to the problem.
  • Exceptions and unhandled promise rejections count as errors.
  • Copy any network request as a cURL command. The Duration and Size columns now show the right values.
  • AI agents that read a replay summary over MCP get the same corrected times.

Read more: Watching a session replay

Team plan drops to $15 per user

  • Pricing

Team is now $15 per user per month, down from $24, with a two-user minimum. Pro stays at $12 and Free is unchanged.

  • Team: $15 per user per month, billed for at least 2 users.
  • Pro: still $12 per user per month.
  • MCP access for AI agents stays free on every plan.

Read more: Pricing

Fixes from our tester rounds

  • Testing
  • Fixes

Outside testers worked through the app and logged what broke. These are the fixes that shipped.

  • Test plans can be archived and restored. A restored plan goes back to the status of its current version.
  • Search boxes treat % and _ as plain characters.
  • A search that fails shows an error with Try again, instead of saying nothing matched.
  • The New plan dialog keeps what you typed if you close it, closes with Esc, and explains a name that is too long in plain words.
  • The plan version you selected stays selected after a refresh.
  • Freezing a plan version with no test cases warns you first.
  • Start Run on an empty suite tells you what to add, instead of creating an empty run.
  • An integration whose connection test failed no longer shows as Connected.
  • A shorter sidebar: Workspace folds into one row, and each project groups its testing pages under Testing.

The CLI, the Playwright reporter and new SDK versions on npm

  • SDKs
  • AI agents
  • Testing

Everything you need to add BugMojo to a React, Next.js or React Native app, or to your CI, now installs from npm.

  • @bugmojo/cli 0.1.0, @bugmojo/playwright-reporter 0.1.0 and @bugmojo/codemap 0.1.0 are published.
  • The Playwright reporter records your test runs in BugMojo with traces, videos and screenshots. A failing test can become a bug with that evidence attached.
  • @bugmojo/widget 0.2.0: a bottom sheet on phones, shake to report, smoother touch replays and device details on every report. @bugmojo/react 0.2.0 includes it.
  • @bugmojo/react-native 0.3.0: Test Mode lets a tester shake the app to run an assigned test case and record the result with evidence.
  • The Claude Code, Codex, Cursor and Gemini CLI plugins now install from our public bugmojo-sdk repository. Installs used to fail for anyone outside our team.
  • Pull requests that BugMojo opens now print verify commands that exist: npx playwright test and npx @bugmojo/cli verify.
  • New help pages for the React SDK, the React Native and Expo SDK, the feedback widget and the CLI.
  • Fixed about 90 broken links between help articles.

Read more: Install the React SDK · Install the React Native and Expo SDK · Record Playwright runs · Install in Claude Code

Test plans, client reports and an inbox for AI agents

  • Testing
  • AI agents

BugMojo now holds your test plans and runs next to your bugs, so the same evidence covers what was tested and what was fixed.

  • Versioned test plans. A published version is frozen, and you can compare any two versions.
  • A review queue for test cases your AI agent proposes: accept, edit or reject each one. Every case is marked automatable or manual, with the reason.
  • Run manual test cases in the browser extension. They record the same evidence as an automated run.
  • Client test reports at a shareable link, with PDF and CSV export and a sign-off tied to the exact report your client read.
  • Tester links let someone outside your workspace run the cases assigned to them, through a link that expires and can be revoked.
  • An Agent Inbox ranks the work waiting for AI agents. An agent claims a task before starting, so two agents never work on the same one.

Read more: Test plans and versions · Client reports and sign-off · The Agent Inbox

No more “can’t reproduce.”

Your testers and customers report a bug in one click. Your developers, and their AI agents, get the replay, the console and network logs, and a failing test. The bug closes only when that test passes.