Running Tests
Once you have test cases ready, you can run them against your AI system and watch the results come in live.
Running a single test
Section titled “Running a single test”Click the play button next to any individual test case. This is useful for:
- Verifying your connection is working.
- Debugging a specific scenario.
- Quick-checking after making changes to your system.
Running all tests
Section titled “Running all tests”Open your platform’s test list and click Run All. Mibo will send every active test case to your AI system and collect the responses.
Watching live logs
Section titled “Watching live logs”As tests execute, you’ll see a live log showing exactly what’s happening:
- Which test case is currently running.
- The input being sent to your system.
- The response coming back.
- Any issues encountered along the way.
The logs update in real time — no need to refresh the page.
Execution statuses
Section titled “Execution statuses”Each execution goes through these stages:
- Waiting — your test run is in the queue, about to start.
- Running — Mibo is actively interacting with your system and collecting responses.
- Completed — all tests have finished. Check the results.
- Error — something went wrong with the connection (this is different from a test failing — an error means Mibo couldn’t reach your system at all).
What happens behind the scenes
Section titled “What happens behind the scenes”When you click “Run,” here’s what Mibo does for each test case:
- Sends the test input to your system, just like a real user would.
- Collects the response and any behind-the-scenes data (trace).
- An AI Judge evaluates the response against your expected behavior.
- Results are saved and your quality dashboard is updated.
Understanding Results Learn how to read your quality scores and failure breakdowns.