Pricing Docs MCP

MCP Guide

Add RunHuman to your AI coding agent and let it handle human QA testing for you.

Installation

claude mcp add --transport http runhuman https://runhuman.com/mcp --header "Authorization: Bearer YOUR_API_KEY"

Replace YOUR_API_KEY with your key from the API Keys dashboard.

Usage Patterns

1. Quick UI/UX Testing

Use when: You need fast feedback on basic functionality.

Example Prompts:

Use RunHuman to test the login form on mysite.com. Try valid credentials, then invalid ones. Tell me if it works and if errors are clear.

Use RunHuman to check if the navigation menu on myapp.com works correctly. Test all links and report any broken ones.

Use RunHuman to verify the signup form validates email addresses properly. Try invalid emails and empty fields.

What happens:

  1. Agent creates test with simple boolean/string schema
  2. Human tester completes test in 2-5 minutes
  3. Agent reports structured results back to you

2. Complex User Flows

Use when: Testing multi-step processes like checkout, onboarding, or account setup.

Example Prompts:

Use RunHuman to test the full checkout flow on staging.myapp.com. Add a product to cart, fill shipping info, and verify the order summary shows correct pricing. Give the tester 10 minutes.

Use RunHuman to test the password reset flow on myapp.com. Request a reset, check if the email arrives, and verify the reset link works. Allow up to 15 minutes since this involves email.

Use RunHuman to test the complete onboarding process on myapp.com. Sign up, verify email, complete the profile setup, and verify the welcome email arrives. Give 15 minutes.

What happens:

  1. Agent sets targetDurationMinutes to 10-15 for longer flows
  2. Human tester works through full flow
  3. Agent extracts structured data about each step
  4. You get detailed breakdown of what worked/failed

Tips:

  • Explicitly mention time needed: “give 10 minutes”, “allow 15 minutes”
  • Break down what to test: “step 1, step 2, step 3”
  • Ask for specifics: “verify email arrives”, “check total is correct”

3. Visual & Accessibility Testing

Use when: Looking for visual bugs, layout issues, or accessibility problems.

Example Prompts:

Use RunHuman to check the mobile navigation menu on mysite.com. Test opening/closing, verify all links work, and report any visual glitches.

Use RunHuman to test the product gallery on myapp.com/product/123. Check if images load, zoom works, and the gallery looks good on mobile. Report any visual issues.

Use RunHuman to test the dashboard on staging.myapp.com for accessibility. Check if screen reader labels are present, keyboard navigation works, and color contrast is sufficient.

Use RunHuman to verify the responsive design on mysite.com. Test on mobile, tablet, and desktop viewports. Report any layout breaks or overflow issues.

What happens:

  1. Agent asks for visual/accessibility checks
  2. Human tester manually inspects UI
  3. Agent gets natural language description of issues
  4. You receive actionable feedback on visual problems

Tips:

  • Mention specific viewports: “test on mobile”, “check tablet view”
  • Ask for screenshots implicitly: “report visual issues”
  • Be specific about what to check: “color contrast”, “keyboard navigation”

Advanced Options

Duration Control

Control how long the tester has to complete the test:

Use RunHuman to test checkout on myapp.com. Give the tester 3 minutes - this should be quick.

Use RunHuman to test the full signup flow. Allow 15 minutes since this involves email verification.

Default: 5 minutes

Extension Control

Prevent tests from running over time:

Use RunHuman to test the login form. Don’t allow duration extension - keep it under 5 minutes.

Use RunHuman to test the checkout flow. Allow duration extension if the tester needs more time.

Default: Extensions allowed

Common Patterns

Before Deploying

Use RunHuman to smoke test staging.myapp.com. Check if the homepage loads, login works, and main navigation functions correctly.

After Fixing a Bug

Use RunHuman to verify the fix for the checkout total bug on staging.myapp.com. Add multiple products and verify the total calculates correctly.

Testing New Features

Use RunHuman to test the new search feature on staging.myapp.com/search. Try different queries, verify results appear, and check if filters work.

Cross-Browser Testing

Use RunHuman to test myapp.com on different browsers. Check if the layout looks correct and all features work across Chrome, Firefox, and Safari.

What Your Agent Does

Behind the scenes, your agent:

  1. Creates test with create_job tool
  2. Waits for completion by calling wait_for_result repeatedly
  3. Extracts results from structured response
  4. Reports back with findings

You don’t need to know the API - just ask naturally.

Tips for Great Results

Be Specific

❌ Bad: “Test myapp.com” ✅ Good: “Test the login flow on myapp.com. Try valid credentials, then invalid password. Check if error messages are clear.”

Provide Context

❌ Bad: “Test checkout” ✅ Good: “Test checkout on staging.myapp.com. Add product to cart, fill shipping, verify total is correct before submitting.”

Set Expectations

❌ Bad: “Test the signup flow” ✅ Good: “Test the signup flow on myapp.com. This involves email verification, so allow 10 minutes.”

Ask for Actionables

❌ Bad: “Check the UI” ✅ Good: “Check the product page UI. Report any visual glitches, broken images, or layout issues.”

Next: See the MCP Reference for technical details on the MCP tools.