Changelog.What changed, week by week.

Everything we ship to qarunbook, what you can now do and why we built it. A new entry every week.

Changes shipped each week, by kind
Week of14 SepWeek of21 SepWeek of28 SepWeek of5 OctLatest
New661214
Improved2486
Fixed2244

2026

Week of 5 October 2026

Latest

Every integration, on every plan

Bugs can now come in from the tools your team already works in, and when one is fixed, the tool it came from hears about it. Every integration and the API are on every plan, Free included. Testing also gets priorities, assignees and owners.

Read the full update

Integrations on every plan

Connect Slack, Microsoft Teams, email, Sentry, PostHog, Jira, GitLab, GitHub, Linear, Zapier or any tool that can send a webhook. A bug raised from any of them lands in the same runbook, and a fix travels back to where it started. Plans differ by how much you use, never by what you can use.

  • Slack
  • Microsoft Teams
  • Sentry
  • PostHog
  • Jira
  • GitLab
  • GitHub
  • Linear
  • Zapier
  • Email
  • Webhooks

Read the integration guides

Sign in from Claude, Cowork and other MCP clients

Connect your AI assistant with OAuth: approve it on a consent page and it's connected, with no token to copy and paste. Each assistant gets its own access, and it reaches only what you can reach.

Connect your AI

Check priorities and triage

Checks have a priority you can set, see as a badge and filter by with 'Urgent & high'. Imported plans keep theirs: P1, Critical or Sev 2 map onto the same four levels. Issues found while testing take a priority and assignees, who are emailed and told in the app.

See priorities in the API

A guide from sign-up to your first test

New accounts answer a few one-tap questions: what you're testing, who's doing it, where the code is and which AI you use. Then a guide says what to do first, what next and why, until you're ready to test, and it follows you into the app. Progress counts what you've actually done, including work your AI assistant or a colleague did, and the welcome emails follow the same path: each one is about your own next step.

Try it

New14

  • Slack: raise any message as a bug with Raise as bug or /qarunbook bug. The bot replies with the issue's code, and says so in the thread when it's fixed.
  • Microsoft Teams: raise a bug from any message, send each app's news to a channel, and hear in the thread when a fix lands.
  • Email to bug: a live app can have its own private address, and mail sent to it becomes an issue. Limit who may send by address or domain.
  • Sentry: new errors that pass your filters become bugs. Resolving in either place resolves both, and a regression reopens the issue.
  • PostHog: raise bugs from PostHog events, and attach a session replay to any issue so testers can watch what happened.
  • Jira: new issues are filed in your Jira project, and moving one to Done marks it fixed and asks for a retest.
  • GitLab: a merged merge request that names an issue's code marks it fixed, on merge or once it's deployed.
  • Incoming webhook: a private URL for each live app that turns a JSON POST from Bugsnag, Rollbar, Datadog or your own monitoring into a bug.
  • GitHub: connect several repositories to one app, each tagged with the platforms it builds, so a fix in one repository marks the bug fixed on its platforms only.
  • Pass notes: record the evidence that a check passed, with files, beside 'Add issue'. Reports show it.
  • Assign a check to the people who should test it. They're emailed, and 'Assigned to me' finds it.
  • Owners for sections and platforms. They hear when checks are added, when a bug is raised in what they own, and when it's fixed.
  • Transfer an app to another workspace, with its checks, results, issues and files.
  • Four new guides: a bug report template, test case examples, a GitHub issue template and a website testing checklist.

Improved6

  • Integrations are a grid of tiles grouped by what they do, and each opens on its step-by-step guide until it's connected.
  • A section's Ref, Check and platform header stays in view while you scroll through its checks.
  • A back-to-top button appears once you're well down a long runbook.
  • Search boxes for checks, issues and people have a clear button, and Escape clears them too.
  • Imported plan sections keep their own names, and plans headed with bare platform codes get readable column names.
  • The MCP server also reads its token from x-api-key, for clients such as claude.ai's custom connector form.

Fixed4

  • Changing someone's role on the team page now changes it on every app they have.
  • GitHub settings say Saved when nothing changed, and give a reason instead of spinning.
  • Sentry setup asks for Organization: Read, which loading your projects needs.
  • A Jira that doesn't sign its webhooks can no longer lock the webhook out.

Read the full update

2026

Week of 28 September 2026

Bug tracking for live apps, and plans from any spreadsheet

An app can now run in live mode, as a bug tracker with priorities, assignees and confirmed fixes. Test plans come in from Excel, CSV or Markdown, or AI reads your own list. GitHub and Linear mark issues fixed when the code lands.

Read the full update

Bug tracking for live apps

Switch an app to live mode and it becomes a bug tracker: issues don't need a check, they take a priority and assignees, and a fix counts once someone confirms it. Share an issue report on screen, as a link, a PDF, Excel or CSV, or by email.

See issues in the API

Import a plan from Excel, CSV or Markdown

Bring the test plan you already have. Upload a spreadsheet or a Markdown file, or let AI read your team's own test list into a runbook you review before anything is saved. Your own check IDs are kept.

Start from a template

Fixes from GitHub and Linear

Every issue gets a code, like QA-12. A merged pull request that names it marks the issue fixed and asks the reporter to retest, or waits until it's deployed. Linear works the same way: Done there means fixed here.

  • GitHub
  • Linear

Read the GitHub guide

The Founding 100

We're picking 100 software teams to get three months of qarunbook Business free, with our help setting up their first test plan.

Read about the Founding 100

New12

  • Paid plans, and a Free plan of one person, one app and a monthly allowance (150 checks tested, 30 issues and 2 AI imports) with every feature on.
  • Reports for every app: on screen, as a PDF, Excel or CSV, as a share link, or by email.
  • A public REST API with workspace keys, documented at /docs.
  • Zapier: triggers for new issues, fixes and results, and actions to record a result, raise an issue and mark it fixed.
  • One Add button for a check, an issue, or many of either from a file. AI reads bugs in bulk, links each to its check and flags ones already open, and nothing is saved until you've reviewed it.
  • Reporters are told in the app and by email when their issue is fixed.
  • A viewer role, and roles per app.
  • A partner program, with a dashboard for partners.
  • Free templates to download as Excel, CSV or Markdown, a software testing checklist, a resources hub and a press kit.
  • New guides: how to write test cases, login and checkout test cases, and retesting vs regression testing.
  • Comparison pages for Zephyr, PractiTest and spreadsheets.
  • Cursor and Codex can install the qarunbook plugin, as Claude Code can.

Improved8

  • App cards show how complete each app is, and how many issues are open.
  • Adding an app is one step: start with AI, a template or an empty runbook.
  • Billing, API keys and Profile open from the sidebar as settings views.
  • Every password field has a show and hide button.
  • Reports and their PDFs show who tested each platform.
  • Failing a check asks why, and the issue starts on the platforms that failed.
  • Free workspaces are emailed when an allowance reaches 80% and when it's used up.
  • The site's top bar has Use cases and Resources menus, so guides and tools are easy to find.

Fixed4

  • Files attached while raising an issue on a live app show straight away, not after a reload.
  • Leaving the report no longer crashes the page.
  • Results and issues recorded over MCP carry the name of the person who owns the token.
  • An imported app's name is checked only against your own workspace.

Read the full update

2026

Week of 21 September 2026

A new look, in light and dark

The whole app has been redesigned, with a proper home page and light and dark mode that follows your device. Issues can carry pictures, videos and files, and the MCP server is clearer for AI assistants.

Read the full update

The MCP server and the Claude Code skill

Your AI assistant can read the test plan, record results and raise issues over MCP. Every tool now has a title and says plainly whether it changes anything. Two new guides cover setting it up in Claude Code and other MCP clients.

Read the Claude Code guide

New6

  • Light and dark mode across the site and the app, following your device unless you choose.
  • Issues can be corrected after they're raised, and carry pictures, videos and files.
  • An owner can put a runbook back to untested.
  • Guides with test plan templates you can copy or download.
  • A free tool that tells you what a test plan forgot.
  • Pages comparing qarunbook with TestRail and Qase, and what free test case management tools really give you.

Improved4

  • A redesigned app and a new home page.
  • New members are invited with a setup link instead of a temporary password.
  • Emails carry the qarunbook logo and match the rest of the product.
  • Workspaces have names, and account pages show when they're loading.

Fixed2

  • Sign-in moved off the home page, and the email paths that could be abused are closed.
  • The sign-in page's title no longer says the site name twice.

Read the full update

2026

Week of 14 September 2026

Workspaces and teams

Each team now has its own workspace. Invite people by email with a role, and give them just the apps they need.

Read the full update

New6

  • Workspaces, members and roles: one runbook per team, not one for everyone.
  • Invite people by email with a role. They bring their own name and password.
  • Give people the apps they need rather than the whole workspace, with a role per app and a link that brings them in.
  • Confirm your email address, and reset your own password.
  • Mark a check N/A over MCP.
  • Search a reference series by its prefix.

Improved2

  • Invitations arrive by email and say the right thing to each person.
  • Refresh sits with the app's own controls, and selected chips look selected.

Fixed2

  • An issue's id is no longer the only thing protecting a runbook.
  • The password reset email now leads to a working page.

Read the full update

Follow along.A new entry every week.

Add the RSS feed to your reader, or create a free account: accounts get product news by email, and you can turn it off under Profile at any time.

The Founding 100.Applications close 30 October.

We're picking 100 software teams to get three months of qarunbook Business free, with our team helping set up the first test plan.

Something missing, or something you'd like next? Write to hello@qarunbook.com.