The mabl MCP server connects your AI client to mabl’s testing tools so you can work with your test data using natural language. Instead of switching between your IDE and the mabl app, use conversational prompts to query test results, investigate failures, and manage your testing workflow directly from your AI client.
For example:
“Show me the latest plan runs.” → “What happened in that failed run?” → “Analyze the failure for me.”
“What applications do I have in mabl?” → “Show me the tests for the Admin Dashboard.” → “Get me the latest results for the login test.”
“What happened with my previous mabl deployment?” → “Show me the test runs from that deployment.” → “Which tests failed?”
Server options
mabl offers two MCP server options: cloud and local. Both connect your AI client to mabl, but they differ in requirements and capabilities:
| Option | Cloud MCP | Local MCP |
|---|---|---|
| Requirements | AI client that supports MCP servers | AI client that supports MCP servers, Node.js, and the mabl CLI |
| Capabilities | Ad hoc cloud execution, querying workspace data, analyzing test failures | All capabilities of the cloud server, local test creation and execution, Playwright export |
| Best for | Users who want a more lightweight, centrally managed way to trigger cloud runs and analyze results | Users who want to create, run, and debug tests locally |