- TRIAGE
- VISUAL
- HISTORY
- HUMAN
Short answer
How can I debug failed tests using my AI?
Your AI uses Mabl tools to fetch specific execution details and visual evidence. It pulls the exact failure reason and screenshots so you can identify the root cause in your chat window. You get a clear breakdown of what went wrong.
The outcomes.
Where your debugging data lands.
The AI organizes the information it pulls from Mabl into these specific outcomes.
TRIAGE
Failure summaries
The AI pulls the specific error message from the execution. You get a plain-text explanation of why the test stopped.
VISUAL
Screenshot analysis
The agent retrieves visual evidence from the failed run. You can see exactly what the UI looked like at the moment of failure.
HISTORY
Execution logs
The AI looks through previous runs to find patterns. This helps you see if a failure is a one-off or a recurring regression.
HUMAN
Manual fixes
The AI identifies the problem but doesn't touch your code. You still need to go into your IDE to apply the actual fix.
The workflow
What your AI does when a test fails.
The AI acts as a bridge between your test runner and your conversation.
Locate the failure
The agent scans your recent Mabl activity to find the specific run that didn't pass.
mb.list_executionsFetch error details
The AI pulls the specific failure reason and metadata for that exact test instance.
mb.get_executionAnalyze visual state
The agent retrieves the screenshots captured during the failed step to show you the UI state.
mb.get_executionSummarize root cause
The AI combines the error text and the visual data to explain the problem to you.
mb.get_execution
Try it
Copy these to start.
Use these prompts to kick off the debugging process.
Starting points
These are opening lines. Swap out the specific execution IDs or test names with your own data.
Mabl Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_GzAfS6LnIq6mGDl04Utu0UzYyZgbblzxeJf9xvxR/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Mabl capabilities are ready to use.
{
"mcpServers": {
"mabl-ai-powered-test-automation-mcp": {
"url": "https://edge.vinkius.com/vk_preview_GzAfS6LnIq6mGDl04Utu0UzYyZgbblzxeJf9xvxR/mcp"
}
}
}List my most recent Mabl executions so I can pick one to debug.
Get the failure reason and screenshots for execution ID 12345.
Why did the 'Checkout Flow' test fail in its last run?
Look at the last failed execution and tell me if it looks like a UI regression.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Mabl once, then ask.
Just follow the link to connect your account. Your credentials stay encrypted and secure. Once connected, you can start asking questions immediately.
Connect Mabl to your AIFAQ
How this task behaves.
- 01
Can the AI fix my Mabl tests automatically?
No. The AI can only read the execution data and screenshots to tell you what happened. You have to perform the actual fix in Mabl.
- 02
Can the AI trigger a new test run?
No. The current tools only allow the AI to list and get details from existing executions.
- 03
Does the AI see all my Mabl data?
The AI only sees the data it needs to answer your specific question, such as execution logs and screenshots.
- 04
Can the AI delete failed executions?
No. The AI has no capability to delete or modify any data within Mabl.
- 05
How does the AI get the screenshots?
The AI uses the execution data provided by Mabl to retrieve the visual evidence associated with a failed step.
More questions about Mabl? See everything the Mabl Connector can do
Connect Mabl to Claude, Cursor, ChatGPT & more
