Release notes··9 updates

Stop runs from chat, multi-touch web gestures, and todos for coding agents

9 updates shipped on August 31, 2026.

  1. Highlight

    Featured this release

    Multi-finger touch gestures for web tests

    Web tests can pinch, rotate, pan and press with up to 10 fingers instead of a single touch point.

    A new touch gesture action covers pinch-to-zoom, multi-finger swipes and rotations, and simultaneous multi-point presses on web apps.

    Describe the gesture in your test steps – pinch, two-finger, multi-touch – and the agent performs it directly instead of substituting a mouse drag.

    More on this: web testing and drag and drop testing.

    # Direct link

  2. Highlight

    Featured this release

    Stop test runs from chat

    Ask the chat agent to pause, stop or cancel a running batch without leaving the conversation.

    The chat agent can now stop an in-progress run the same way the results page's Stop button does. Say "stop", "pause", "cancel" or "abort" and it cancels whatever remains – tests that already finished keep their results.

    Works on the run in your current chat session, or on any run when you point it at a results URL.

    More on this: web testing.

    # Direct link

  3. Update 03 of 09

    Turn chat attachments into File Upload test data

    Files you share with chat, or ask it to generate, can be saved straight into a reusable File Upload config.

    Attach a file to a conversation, or ask the agent to generate one, and it can persist it as a File Upload config for your test cases – no separate manual upload step first.

    More on this: file upload and export testing.

    # Direct link

  4. Update 04 of 09

    Ask the chat agent who created or last edited something

    Test plans, test cases, scenarios, configs and device presets can be filtered by creator or last editor.

    Ask for "the test plan Alex last edited" and chat narrows the list instead of returning everything. Test plans, test cases, test scenarios, configs and device presets can all be filtered by who created them or who updated them last, and the agent resolves a name to the right teammate before filtering.

    More on this: how teams run QA.tech and QA teams.

    # Direct link

  5. Update 05 of 09

    Clearer errors for incompatible iOS builds

    Device-only iOS builds are rejected at upload time with an inline explanation and a link to supported formats.

    QA.tech now checks iOS app builds when you upload them and flags device or App Store builds that cannot run in the simulator, showing the reason inline under the file field with a "Learn more" link – so you find out at upload instead of mid-run.

    More on this: mobile testing and mobile app testing.

    # Direct link

  6. Update 06 of 09

    Follow-up PR reviews keep their evidence

    A follow-up review carries forward the original reproduction evidence and supporting tests.

    The sticky pull request review comment now keeps full verification – the reproduction plus the tests behind the verdict – through every round of review, including rounds where the latest commit did not change the outcome.

    More on this: pull request testing and PR and CI/CD testing.

    # Direct link

  7. Update 07 of 09

    Manage labels and cross-tool references from chat

    Create, assign and filter test case labels, and link Linear, Jira or GitHub references, from the conversation.

    The chat agent can now create and manage project labels, assign or unassign them on test cases (including in bulk) and filter results by label name. It can also link external references – Linear issues, GitHub pull requests, Jira tickets, TestRail cases or Xray tests – directly to a test case.

    Labels now render as neutral chips with a colored dot instead of solid fills, and the project's application shows as an inline chip alongside labels and references. Past four labels or references, the extras collapse into a "+N" chip you can hover to expand.

    More on this: the Jira, Linear and TestRail integrations.

    # Direct link

  8. Update 08 of 09

    MCP tool docs now cover the full toolset

    Every tool the QA.tech MCP server exposes is documented, grouped by area with read and write scopes.

    The MCP integration page now documents every available tool – runs, test cases, issues, knowledge base, rules, feature search, chat, release checks and remote tunnels – grouped by area with accurate read and write scope descriptions, so AI coding assistants (and you) know exactly what is callable.

    More on this: the QA.tech MCP server and the MCP integration.

    # Direct link

  9. Update 09 of 09

    See a run network path in Runtime, or ask chat about it

    Outbound IP, proxy and SSH tunnel details now appear on test results and can be answered in chat.

    The Runtime section on a test result now shows the connectivity details that applied to that run – Outbound IP, Proxy, SSH Tunnel – with a "More details" view for the full picture.

    You can also ask the chat agent about a run's network path directly.

    More on this: IP access control and multi-environment testing.

    # Direct link

Your team moves fast. Can your testing keep up?

QA.tech agents test your product autonomously, so moving fast never means shipping broken. See how it works in a 30-minute demo.

Get a demo