Automations, on-demand PR reviews, and compactable conversations

Highlight
Featured this release
Automations: schedule test runs and chat workflows
Project Settings now runs your chat workflows or test plans on a schedule, via the API, or on demand.
Create an automation under Project Settings → Automations for an AI Assistant (chat) action or a Run Test Plan action, then set it to fire hourly, daily, weekly, monthly or once, in your project's timezone.

Run Now fires immediately, even while an automation is paused, and any automation can be triggered from the public API. A nested History view links each run back to its conversation or test result.

More on this: smoke testing and automated regression testing.
# Direct link
Update 02 of 05
Chat agent can research sibling repositories
The codebase tool can look at any repository your GitHub connection can reach.
Mention a sibling repository by name or paste its GitHub URL and the chat agent's codebase tool explores that repository instead of only the one mapped to your project. The agent run card shows which repository it looked at.

Code access still needs to be enabled on a GitHub mapping, and that mapping's installation determines which repositories are reachable.

More on this: the GitHub integration and AI test generation.
# Direct link
Update 03 of 05
Chat conversations report status and linked runs
Conversations started over the API or MCP expose a status field and their linked runs.
Conversations started through the public API or MCP now return a status – in progress, completed, failed or cancelled – along with a list of runs carrying each run's short ID, status and result.

Poll the conversation until its status is terminal, then fetch the run to see the outcome, instead of guessing when a result is ready.

More on this: the QA.tech MCP server and the CI/CD API integration.
# Direct link
Update 04 of 05
Compact long conversations with /compact
A /compact command condenses a long conversation, with a context-usage indicator in the composer.
Type /compact in the chat composer to condense a long conversation into a fresh, smaller context, optionally with a short note describing what to keep.

A context-usage indicator in the composer shows how much of the conversation window you have used, and compacted conversations persist so the summary survives a refresh.

More on this: what QA.tech does.
# Direct link
Update 05 of 05
Test editor remembers your Environment and Device
Manual Environment and Device selections in the test editor stick between visits.
Pick an Environment or Device while editing a test and QA.tech remembers it next time you open that project's editor. Opening a test from a specific run or test plan applies that run's options for the visit without overwriting your saved defaults, and override chips now show the environment name rather than a raw ID.

More on this: multi-environment testing and mobile testing.
# Direct link

Automations: schedule test runs and chat workflows

Chat agent can research sibling repositories

Chat conversations report status and linked runs

Compact long conversations with /compact

Test editor remembers your Environment and Device

Your team moves fast. Can your testing keep up?