Skip to main content

Bundle Studio

Bundle Studio is an interactive testing environment built into your MCPBundles dashboard. It lets you execute tools directly in your browser, see real-time results, and get help from an AI assistant—all with proper credential context.

What is Bundle Studio?​

Think of Bundle Studio as a playground for your bundle's tools. Instead of testing tools by asking your AI to call them, you can:

  • Execute tools directly with a simple form interface
  • See exactly what inputs a tool needs and what outputs it returns
  • Test tools before adding them to your AI workflow
  • Debug tool execution issues in a controlled environment
  • Get guidance from an AI assistant that understands your bundle

Why Use Bundle Studio?​

Before Bundle Studio:

  • Add bundle to your AI → hope it works → debug in production
  • Wonder what inputs a tool actually needs
  • Guess at credential issues when tools fail

With Bundle Studio:

  • Test tools interactively before deploying
  • See exactly what each tool does with real data
  • Verify credentials are working correctly
  • Learn tool behavior in a safe environment

When to Use Bundle Studio​

Bundle Studio is perfect for:

Testing New Tools​

Before adding a bundle to your AI, test its tools to understand what they do and how they work.

Debugging Issues​

When a tool isn't working in your AI, use Studio to isolate the problem and see detailed error messages.

Learning Tool Behavior​

Explore what inputs a tool accepts, what outputs it returns, and how it handles different scenarios.

Verifying Credentials​

After connecting a provider, test tools that use it to confirm your credentials are working correctly.

Exploring Tool Capabilities​

Discover what a tool can do before building workflows around it.

Accessing Bundle Studio​

There are two ways to enter Bundle Studio:

From the Bundle Page​

  1. Go to Dashboard → Bundles
  2. Click on any bundle you've enabled
  3. Click the "Open Studio" button (top right of the assistant card)
  4. Studio opens with all bundle tools available

URL format: https://www.mcpbundles.com/dashboard/bundles/[bundle-id]/studio

From a Tool Page​

  1. Go to Dashboard → Tools
  2. Click on any tool
  3. Find the "Bundles" section
  4. Click "Execute in [Bundle Name]" on any bundle
  5. Studio opens with that tool focused and expanded

URL format: https://www.mcpbundles.com/dashboard/bundles/[bundle-id]/studio?tool=[tool-id]

Bundle Must Be Ready

Bundle Studio only works for bundles that are Operational (all required providers connected and verified). If a bundle shows "Blocked" or "Partial" readiness, connect its providers first.

The Studio Interface​

Bundle Studio has two main sections:

Left Panel: Tool Execution​

A collapsible panel showing all tools in the bundle:

📦 Bundle: Marketing Automation
├── ▶ list_campaigns
├── ▼ get_campaign_stats    ← Expanded
│   ├── [Input Form]
│   ├── [Run Tool Button]
│   └── [Results Display]
└── ▶ create_contact

For each tool:

  • Tool name and description badge
  • Status badge showing if it's ready to use
  • Expand/collapse arrow
  • When expanded:
    • Auto-generated form based on the tool's input schema
    • "Run Tool" button (disabled until required fields are filled)
    • Results section showing execution output

Right Panel: AI Assistant​

A chat interface with an AI assistant that:

  • Understands your bundle and its tools
  • Can help you decide what to test
  • Explains tool inputs and outputs
  • Suggests test scenarios
  • Guides you through troubleshooting

Chat persists across page navigation - if you switch bundles or leave Studio, your conversation stays intact when you return.

Collapse Controls​

  • "Hide Tools" button (top-left) collapses the tool panel for more chat space
  • "Show Tools" button brings it back
  • Chat is always visible

Using Bundle Studio​

Step-by-Step Execution​

1. Select a Tool

  • Browse the tool list in the left panel
  • Click on any tool name to expand it

2. Fill the Form

  • Studio generates input fields based on the tool's schema
  • Required fields are marked and must be filled
  • Field types:
    • Text inputs for strings
    • Number inputs for integers/floats
    • Checkboxes for booleans
    • Dropdowns for enums
    • Text areas for long text or JSON objects

3. Run the Tool

  • The "Run Tool" button is disabled until all required fields are filled
  • Once ready, click "Run Tool"
  • Button shows "Running..." with a spinner while executing

4. View Results

  • Results appear below the form
  • Success:
    • Green checkmark with execution time
    • Response text (if the tool returns text)
    • Result data (formatted JSON)
  • Failure:
    • Red X with error details
    • Full error message for debugging

Example: Testing a Campaign Tool​

Let's test the get_campaign_stats tool:

Step 1: Expand the tool

▼ get_campaign_stats
  "Fetch statistics for a specific email campaign"

Step 2: Studio generates form from schema

Campaign ID* (required)
[___________________]  ← Text input

Include Detailed Metrics
[ ] ← Checkbox (optional)

Step 3: Fill required field

Campaign ID*
[campaign-123-abc]  ✓

Step 4: Click "Run Tool"

[Running...]

Result (847ms)
{
  "campaign_name": "Summer Promotion 2024",
  "total_sent": 1250,
  "opens": 412,
  "open_rate": 0.3296,
  "clicks": 89,
  "replies": 23
}

Form Field Types​

Bundle Studio generates forms automatically from tool input schemas. Here's what to expect:

Text Input​

For: string type fields

Email Address*
[user@example.com]

Number Input​

For: integer or number type fields

Campaign ID*
[12345] ← Only accepts numbers

Checkbox​

For: boolean type fields

Include Details
[✓] ← Checked = true

For: Fields with enum values

Status
[Active ▼]
  - Active
  - Paused
  - Completed

Text Area​

For: Long strings or object fields

Filter Criteria (JSON)
┌─────────────────────────┐
│ {                       │
│   "status": "active",   │
│   "date_after": "2024"  │
│ }                       │
└─────────────────────────┘

Array Input​

For: Array fields

Tags*
[tag1, tag2, tag3]
Schema-Driven Forms

Forms are auto-generated from the tool's JSON Schema. If a field looks wrong, the tool's schema might need updating (contact the bundle author).

Understanding Results​

Success Response​

When a tool executes successfully, you'll see the results:

If the tool returns a text response:

Response (847ms)
Campaign "Summer Promotion" has 412 opens from 1,250 sent emails.
Open rate: 32.96%

Top performing email subject lines:
• "Limited Time Offer" - 45% open rate
• "Your Exclusive Discount" - 38% open rate
• "Don't Miss Out" - 31% open rate

Result data (always shown):

Result (847ms)
{
  "campaign_id": "abc123",
  "campaign_name": "Summer Promotion",
  "metrics": {
    "sent": 1250,
    "opens": 412,
    "open_rate": 0.3296,
    "clicks": 89,
    "replies": 23
  }
}

The result data shows the actual JSON response from the tool execution.

Error Response​

When a tool fails, you'll see a clear error message:

Header:

✗ Execution Failed

Error Message:

Campaign not found

The campaign ID "invalid-123" does not exist in your Smartlead account.

Please verify:
• The campaign ID is correct
• The campaign hasn't been deleted
• You're using the right Smartlead account

Common error types:

  • Authentication errors - Credential issue (check provider connection)
  • Validation errors - Invalid input format (check field requirements)
  • API errors - Provider service issue (check provider status)
  • Not found errors - Resource doesn't exist (check IDs/names)

Working with the AI Assistant​

The Studio AI assistant can help you test tools effectively.

Example Conversations​

Understanding a tool:

You: What does the create_lead tool do?

AI: The create_lead tool adds a new lead to your Smartlead campaign.
It needs:
- email (required)
- first_name (optional)
- last_name (optional)
- campaign_id (required)

Would you like me to help you test it?

Getting help with inputs:

You: I want to test get_campaign_stats but I don't know my campaign ID

AI: You can use the list_campaigns tool first to see all your campaigns
and their IDs. Would you like me to walk you through that?

Debugging an error:

You: The create_contact tool failed with "invalid email format"

AI: The tool requires a valid email address. Make sure:
- Email has @ symbol
- Domain is valid (e.g., user@example.com)
- No spaces or special characters

Try again with a properly formatted email.

Tips for Using the Assistant​

  • Ask specific questions about tools, inputs, or errors
  • Request test scenarios if you're unsure what to try
  • Get schema explanations when field requirements are unclear
  • Troubleshoot failures by sharing error messages

Troubleshooting​

"Run Tool" Button is Disabled​

Cause: Required fields are not filled

Solution:

  1. Look for fields marked with * (asterisk)
  2. Fill all required fields
  3. Button enables automatically when ready

Tool Execution Fails​

Cause 1: Credential Issues

Error: Authentication failed - Invalid API key

Solution:

  1. Go back to the bundle page
  2. Click on the provider mentioned in the error
  3. Verify credential status
  4. Re-run validation tool
  5. If needed, reconnect the provider

Cause 2: Invalid Input

Error: Validation failed - campaign_id must be a string

Solution:

  1. Check the field type in the form
  2. Ensure input matches expected format
  3. For JSON fields, validate JSON syntax
  4. Check tool documentation for examples

Cause 3: API Service Error

Error: 503 Service Unavailable - Provider is down

Solution:

  1. Check provider's status page
  2. Wait and retry in a few minutes
  3. If persistent, contact support

Tool Not Appearing in Studio​

Cause: Bundle is not operational

Solution:

  1. Go to bundle page
  2. Check "Bundle Readiness" section
  3. Connect any missing providers
  4. Verify all providers show green checkmark
  5. Return to Studio - tool should now appear

Results Look Wrong​

Cause: Tool might have bugs or unexpected behavior

Solution:

  1. Try different input values
  2. Check tool's expected output schema
  3. Compare with tool description
  4. If consistently wrong, report to bundle author

Best Practices​

Start Simple​

Begin with read-only tools (list, get, search) before testing write operations (create, update, delete).

Test with Real Data​

Use actual IDs and values from your accounts for realistic results.

Save Test Inputs​

If you'll test frequently, keep a note of test values (campaign IDs, contact names, etc.)

Use the Assistant​

Don't guess - ask the AI assistant if you're unsure about anything.

Test Before Deploying​

Always test new tools in Studio before adding bundles to production AI workflows.

Check Credentials First​

Verify provider connections are working before blaming the tool.

Limitations​

No Persistence​

Studio doesn't save form inputs between sessions. You'll need to re-enter values each time.

Bundle Context Only​

Tools execute in the context of the specific bundle. If a tool is in multiple bundles, you might see different behavior depending on which bundle's Studio you use (based on credential bindings).

Execution Quotas​

Tool executions in Studio count toward your account usage limits, just like executions from your AI.

No Batch Execution​

You can only run one tool at a time. For testing multiple tools in sequence, use the AI assistant in Studio or your AI client.

What's Next?​

Once you've tested your tools in Bundle Studio and confirmed they work:

1. Add to Your AI​

Follow the Quick Start Guide to connect your bundle to ChatGPT, Claude, Cursor, or another AI.

2. Build Workflows​

Combine tools into powerful AI workflows. See Understanding Bundles for ideas.

3. Share with Team​

If your bundle is published, share it with colleagues who can add it to their accounts.

4. Create Custom Bundles​

Build your own bundles with exactly the tools you need. See Understanding Bundles - Creating a Bundle.

Need Help?​

  • Email: help@mcpbundles.com
  • Docs: Browse the sidebar for more guides
  • In-app help: Use the AI assistant in Bundle Studio

Bundle Studio makes testing tools fast, safe, and easy. Happy testing!