July was an enterprise month. Most of what shipped answers the questions that come up once a test team grows past a handful of people: how do we hand documentation to an auditor, where is the evidence for this step, who changed this result and when, and how do we run the same suite against a feature branch without rebuilding the integration every time.
Alongside that, Katalon AI Assistant connects to one more system your team already lives in, Production Insights got more honest about what it captured, and Katalon Studio 11.4.0 made execution and reporting faster to work with. Here's everything that shipped.
July's test management work is about getting information out of the platform in a form other people accept, and keeping a defensible record of what happened.
Export formatted test case documents in bulk, including test case ID, steps, linkages, and all additional information. This is the answer to the recurring request from regulated teams and anyone who has to hand a reviewer, an auditor, or a client something readable outside the tool.
Attach screenshots, documents, or other evidence directly to an individual test step, so reviewers and testers find the reference where they need it rather than three scrolls away in a description field. Images render as a thumbnail on the step. PDFs, Word documents, and CSVs appear as a card with a preview, which you can open or download.
The limits, so you can plan around them:
| Step attachment limits What you can attach to a single manual test step. |
|
|---|---|
| IMAGE FILES | PNG, JPG, JPEG, GIF, WebP. Uploaded images appear as a thumbnail on the step. |
| DOCUMENT FILES | PDF, DOCX, XLSX, CSV, TXT, MD. These appear as a card with a preview thumbnail you can open or download. |
| MAXIMUM FILE SIZE | 20 MB per file. |
| MAXIMUM FILES | 5 files per step. |
An immutable, read-only log of every create, modify, approve, and delete action on test cases, executions, and results. If you have ever been asked to prove that a result was not edited after the fact, this is the record that answers it.
Integrate your base GitHub repository once, then pick any target branch from a dropdown at execution time. No per-branch setup for every feature branch you want to run against. To keep the workspace readable, test management only syncs tests from your integrated branch, which defaults to main if you integrate the base repository and change nothing else. One thing to watch: the test suite you select in the UI has to exist on the target execution branch, or the run will not find it.
Connect it through OAuth, alongside the existing Google Drive and Confluence connectors, then use connected work item, Test Plan, document, and page context to inform AI-assisted testing. You can switch a connector's participation per chat without disconnecting the account, which matters when you want a clean context for one task and the full set for another.
Generated tests retain more acceptance context and sample data, so there is less filling in blanks after generation. Review output is grouped by feature and ordered Happy, Alternative, Negative, and Edge, which means you can review a feature's coverage as a set and see immediately where the negative and edge cases thin out.
Production Insights spent July making its picture of real user behavior more complete and easier to search.
Switch between most-visited and least-visited pages to see both ends of the distribution. The top end tells you where to prioritize automation coverage. The bottom end is the more interesting half: low-traffic pages are candidates for cleanup, deprioritization, or a conversation about whether they should exist at all.
The agent now captures programmatic input changes, interactions inside same-origin iframes, and file input events more reliably. If your app sets field values in code or wraps part of the flow in an iframe, the recorded session and the generated test reflect more of what actually happened.
Search User Sessions by the number of recorded steps, so you can find sessions at the right level of detail for review and generation instead of opening a dozen 3-step sessions looking for a substantial one.
Run multiple test cases in a suite concurrently, within the suite's configured concurrency threshold. Parallelism now goes a level deeper than the suite.
The new consolidated HTML report combines data binding and retry attempts into a single file, so reviewing results follows the execution flow instead of jumping between artifacts. You can also split a test suite HTML report into smaller per test case reports, which helps when a large suite report is slow to open.
Interactive Mode is now generally available, local device streaming stays live during a session, notification interactions are recordable, and recorder optimization settings moved into Project Settings.
Follow an expandable plan and progress timeline, watch responses stream as they are generated, and see the reasoning behind each task. When an agent takes a wrong turn, you can see where it happened instead of inferring it from the result.
Clearer variable-binding states, plus fixes across recording, API testing, licensing, reporting, desktop testing, and execution.
For a look at how these changes work in practice, including Interactive Mode in the Smart Mobile Recorder, watch the latest What's New in Katalon Studio video:
Download Studio 11.4.0 · See the release notes
If June was about shortening the distance between a real user session and a working test, July was about what happens after that test exists and other people need to see it.
Documentation exports in a format an auditor accepts. Evidence attaches to the step it belongs to. An immutable log answers who changed what. Branch selection stops being setup work.
Katalon AI Assistant reads one more system your team already uses and hands back coverage you can review as a set. Studio 11.4.0 runs deeper in parallel and produces reports that match how you actually read them.
Try what's new, and tell us what we should change next.