Streaming Thunders Copilot thinking, API response inspector, and test set pages
TL;DR
Thunders Copilot now streams its thinking in real time, API steps get a full response inspector with one-click variable extraction and assertions, and every test set gets its own page.
FEATURE
- Streaming Thunders Copilot thinking: watch Thunders Copilot stream its thinking and actions in real time instead of waiting behind a loading spinner.
- API response inspector: open an API step's full response after execution, navigate the JSON, and extract any value as a reusable variable in one click.
- API response assertions: validate status codes, headers, or body values directly as part of the test flow.
- Dedicated test set pages: every test set gets its own page with pass rates, set runs, and its test cases, all searchable and filterable in one place.
- Multi-tab preview: see every tab your test opens in the preview header, each with its page title and favicon, during live execution and in the run view.
- Editor keyboard shortcuts: common test case editor actions now have hotkeys, with tooltip hints to help you learn them.
IMPROVEMENT
- Accessibility "Always" severity: a new "Always" severity level in the accessibility persona for finer control over which issues get flagged.
- Console-exception detection on by default: now enabled by default for the QA Engineer persona.
- Bulk run actions: re-execute or stop test runs in bulk from the runs list.
- Longer variable values: variable values can now hold up to 2048 characters, up from 512.
- STL test assets: upload STL 3D-model files as test assets.
- Cleaner environment dropdown: the environment dropdown no longer shows the URL, for a tidier look.
- Thunders MCP: listing test cases now returns their labels and test sets.
FIX
- Retry counting: CI run reports now count only the final retry outcome per test, so retries are no longer double-counted.
- CSV retry fields: re-executing CSV dataset runs now preserves batch fields.
Ready to ship faster with smarter testing?
