Testdino MCP

What testdino-hq/testdino-mcp does

testdino-hq/testdino-mcp gives MCP-compatible AI clients access to TestDino test execution and test-management data. It can inspect organizations and projects, search automated test runs, retrieve suite and case details, and expose execution evidence such as logs, traces, screenshots, videos, and visual-diff images.

The server also covers manual QA workflows. Agents can search, create, and update manual test cases; browse or create nested suites; manage manual runs; and create, inspect, or update releases. Release tools support milestone-style organization, nested releases up to three levels, linked issues, dates, environments, branches, testers, and rolled-up progress data.

How it works

The server uses a TestDino personal access token supplied through the MCP client's environment configuration. Start with health to validate the connection and obtain account, organization, project, and role information. Other tools generally use the resulting organization or project identifiers.

For failure investigation, get_debug_evidence combines the flake verdict, regression boundary, and temporary artifact links in one response. get_testcase_details provides execution-level evidence, while debug_testcase aggregates historical runs and returns data intended for AI-assisted root-cause analysis. verify_fix compares runs after a baseline against the run where a proposed fix was identified.

The audit flow is specifically for TestDino Playwright audits. First retrieve audit context, analyze local Playwright code using that context, and then submit the completed report. The server does not require raw source files to be uploaded for this flow.

Setup and configuration

Install through an MCP client's standard server configuration using the npm package:

{
  "mcpServers": {
    "TestDino": {
      "command": "npx",
      "args": ["-y", "@testdino/mcp"],
      "env": {
        "TESTDINO_PAT": "your-pat-here"
      }
    }
  }
}

Replace the placeholder with a PAT created in TestDino under User Settings and Personal Access Tokens. Cursor uses ~/.cursor/mcp.json; equivalent MCP server configuration can be used in other compatible clients. Restart the client after adding the entry, then ask the agent to check TestDino health.

Tools and capabilities

• Filter automated runs by branch, status, time interval, author, environment, tags, commit text, or run counter, with pagination.
• Retrieve run statistics, suites, cases, Git metadata, pull-request information, and error details; batch run-detail requests support up to 20 IDs.
• Search cases across selected runs or matching runs, then inspect steps, code locations, console output, errors, and stored artifacts.
• Classify retry behavior as deterministic, flaky, or inconclusive, identify the last passing and first failing runs, and compare post-fix history.
• Retrieve audit context or saved reports and submit completed Playwright audit findings.
• Manage manual cases, suites, runs, releases, automated-test links, comments, and linked Jira issue references where supported by the relevant tool.

Limitations and notes

Role information from health is informational and should not be treated as a security guarantee. AI insights and AI fixes may be generated lazily; when a response reports processing, poll get_ai_insights rather than repeatedly requesting the original analysis. Artifact download links are short-lived and should be used promptly.

A flake verdict needs retries enabled; a single attempt is inconclusive. Test-case names are only sufficient when paired with a run identifier because names can repeat across runs. The TestDino audit flow applies to Playwright automated tests and must be explicitly requested by naming a TestDino audit. Write operations require suitable permissions, and automated-test linking also depends on the applicable plan feature.