ai/studio

Building Minds in Quantum Minds

Introduction

This guide walks you through the process of creating, configuring, and deploying AI-powered minds in Quantum Minds. Whether you're starting from scratch, using a template, or generating a mind using AI, this document provides comprehensive guidance on building effective mind workflows.

Mind Creation Methods

Quantum Minds offers four different approaches to create minds, each suited to different needs and expertise levels:

1. Starting from Scratch (Blank Template)

Best for: Custom workflows with specific requirements, experienced users who want full control

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Select "Blank Template"
  4. Enter a mind title and description
  5. Click "Create" to open the Mind Builder interface
  6. Add and connect operators on the canvas
  7. Configure each operator's parameters
  8. Save and test your mind

2. Using Pre-built Templates

Best for: Common use cases, beginners, or when you need a starting point for customization

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Browse available templates in the Templates tab
  4. Select a template that matches your needs
  5. Review the template structure and description
  6. Click "Create" to create a copy of the template
  7. Customize as needed
  8. Save and test your mind

Available template categories include:

3. AI-Assisted Creation

Best for: Quick mind creation, users who know what they want but aren't sure how to build it

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Select "Create Mind using AI"
  4. Enter a detailed description of what you want the mind to do
  5. Specify data sources or collections to use (optional)
  6. Select an AI model for generation
  7. Click "Create" to generate the mind
  8. Review and refine the generated mind
  9. Save and test your mind

Tips for effective AI-assisted creation:

4. Importing Mind Schema

Best for: Reusing or sharing minds, migrating minds between environments

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Select "Import MindSchema"
  4. Upload a previously exported mind schema file (.json)
  5. Review imported mind details
  6. Click "Create" to finalize import
  7. Update connections or settings if needed
  8. Save and test your mind

Mind Builder Interface

Canvas

The main workspace where you build your mind:

Toolbar

Located at the top of the builder, providing essential tools:

Operator Panel

Located on the left side, providing access to operators:

Properties Panel

Located on the right side when an operator is selected:

Adding and Configuring Operators

Adding Operators to the Canvas

  1. Browse or search for an operator in the Operator Panel
  2. Drag the operator onto the canvas, or click to add it at the default position
  3. Position the operator where you want it in your workflow

Configuring Operator Parameters

  1. Select an operator on the canvas
  2. The Properties Panel will display available parameters
  3. Fill in required parameters (marked with *)
  4. Configure optional parameters as needed
  5. Use the documentation tab for guidance on parameter usage

Working with Variables

Every operator creates output variables that can be referenced by other operators:

  1. Variable naming follows the pattern: $OperatorName_ID.output.property
    • Example: $TextToSQL_001.output.content
  2. To use a variable:
    • Select the target operator
    • In the input field where you want to use the variable, type `# Building Minds in Quantum Minds

Introduction

This guide walks you through the process of creating, configuring, and deploying AI-powered minds in Quantum Minds. Whether you're starting from scratch, using a template, or generating a mind using AI, this document provides comprehensive guidance on building effective mind workflows.

Mind Creation Methods

Quantum Minds offers four different approaches to create minds, each suited to different needs and expertise levels:

1. Starting from Scratch (Blank Template)

Best for: Custom workflows with specific requirements, experienced users who want full control

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Select "Blank Template"
  4. Enter a mind title and description
  5. Click "Create" to open the Mind Builder interface
  6. Add and connect operators on the canvas
  7. Configure each operator's parameters
  8. Save and test your mind

2. Using Pre-built Templates

Best for: Common use cases, beginners, or when you need a starting point for customization

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Browse available templates in the Templates tab
  4. Select a template that matches your needs
  5. Review the template structure and description
  6. Click "Create" to create a copy of the template
  7. Customize as needed
  8. Save and test your mind

Available template categories include:

3. AI-Assisted Creation

Best for: Quick mind creation, users who know what they want but aren't sure how to build it

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Select "Create Mind using AI"
  4. Enter a detailed description of what you want the mind to do
  5. Specify data sources or collections to use (optional)
  6. Select an AI model for generation
  7. Click "Create" to generate the mind
  8. Review and refine the generated mind
  9. Save and test your mind

Tips for effective AI-assisted creation:

4. Importing Mind Schema

Best for: Reusing or sharing minds, migrating minds between environments

Process:

  1. Navigate to the Quantum Minds dashboard
  2. Click "Create New Mind"
  3. Select "Import MindSchema"
  4. Upload a previously exported mind schema file (.json)
  5. Review imported mind details
  6. Click "Create" to finalize import
  7. Update connections or settings if needed
  8. Save and test your mind

Mind Builder Interface

Canvas

The main workspace where you build your mind:

Toolbar

Located at the top of the builder, providing essential tools:

to see available variables

Dynamic Parameters

Some operators support dynamic parameters that can change based on context:

  1. Session variables: ${session.variable_name}
  2. User variables: ${user.variable_name}
  3. Environment variables: ${env.variable_name}
  4. Trigger variables: ${trigger.variable_name}

These can be used to create dynamic workflows that adapt to different users, sessions, or environments.

Creating Connections

Basic Connections

To create a linear flow between operators:

  1. Hover over the output connector of the source operator
  2. Click and drag to the input connector of the target operator
  3. Release to create the connection
  4. The connection will show as a line between the operators

Managing Multiple Outputs

When an operator has multiple outputs (like Flow.Condition):

  1. Each output appears as a separate connector point
  2. Connect each output to the appropriate next operator
  3. Connections are labeled according to their output type
  4. Different styling indicates different types of connections (normal flow, conditional path, etc.)

Auto-Connecting Variables

When you reference a variable from one operator in another:

  1. The system automatically creates a visual connection between those operators
  2. These connections appear as dotted lines to distinguish them from explicit connections
  3. You can hide or show these auto-connections using the view settings

Testing and Debugging

Running a Mind

To test your mind:

  1. Click the "Run" button in the toolbar
  2. The system will execute operators in the correct sequence based on connections
  3. A progress indicator shows which operators are currently executing
  4. Results appear in the output panel at the bottom of the screen

Debugging Mode

For troubleshooting complex minds:

  1. Enable debugging mode from the toolbar
  2. Execution will pause after each operator
  3. Review the outputs before continuing
  4. Use the step buttons to control execution
  5. View detailed logs in the debug panel

Common Issues and Solutions

Issue Possible Causes Solutions
Missing variable reference Typo in variable name, Operator not yet executed Check variable spelling, Ensure connected operators execute in the correct order
Empty results No data returned, Filter too restrictive Verify data source has content, Check filter conditions
Timeout errors Operation too complex, External service delay Optimize queries, Break into smaller operations, Check service status
Out of memory Too much data, Infinite loops Limit data size, Add pagination, Check for circular references
Incorrect output format Mismatched types, Parsing errors Verify input/output compatibility, Add type conversion

Advanced Mind Building

Creating Reusable Patterns

For efficiency in building multiple similar minds:

  1. Identify common workflows in your organization
  2. Create mind templates for these patterns
  3. Export mind fragments as separate schema files
  4. Share these patterns with your team
  5. Import patterns into new minds to avoid rebuilding common elements

Working with Models

When using operators that leverage AI models:

  1. Select appropriate models for your task in the operator properties
  2. Consider model strengths and limitations:
    • Larger models (70B+) for complex reasoning
    • Smaller models (7B) for faster, simpler tasks
    • Specialized models for specific domains
  3. Test different models to find the best balance of performance and accuracy
  4. Consider cost implications of different model choices

Performance Optimization

For minds that need to process large amounts of data or run frequently:

  1. Filter data early in the workflow
  2. Use efficient operators for heavy processing
  3. Consider chunking large datasets
  4. Use database-native operations when possible
  5. Cache results where appropriate
  6. Monitor execution times to identify bottlenecks

Mind Templates

Templates provide ready-to-use starting points for common use cases. Here are some popular templates:

Data Analysis Templates

Document Processing Templates

Industry-Specific Templates

Deploying Minds

Execution Methods

Once your mind is built, you can deploy it in several ways:

  1. On-Demand Execution:

    • Run the mind manually when needed
    • Use for ad-hoc analysis or occasional tasks

  2. Scheduled Execution:

    • Set up recurring schedules (hourly, daily, weekly, monthly)
    • Use for regular reporting or data processing
    • Configure notification options for results

  3. Event-Triggered Execution:

    • Configure the mind to run in response to events
    • Events can include new data arrival, API calls, or system alerts
    • Use for real-time processing and responses

  4. API Endpoint:

    • Expose the mind as an API endpoint
    • Call from external applications
    • Pass parameters to customize execution

Publishing as an App

To make minds accessible to users across your organization:

  1. Finalize and test your mind thoroughly
  2. Click "Publish" from the mind details page
  3. Configure publishing settings:
    • App name and description
    • Access permissions
    • Input parameters that users can modify
    • Refresh settings
    • Visualization preferences
  4. Generate a MINDSHARE_KEY for secure access
  5. Use the published link to share the app
  6. Embed in dashboards or other applications if needed

Access Control

Manage who can access and use your minds:

  1. User-Level Permissions:

    • View: See mind results
    • Execute: Run the mind
    • Edit: Modify the mind
    • Manage: Change settings and permissions

  2. Role-Based Access Control:

    • Assign permissions to roles rather than individuals
    • Manage access at scale
    • Integrate with your organization's identity management

  3. Collection-Level Security:

    • Control access to underlying data sources
    • Ensure data governance compliance
    • Implement row-level security where needed

Best Practices

Mind Organization

User Experience

Maintenance

Examples and Scenarios

Example 1: Customer Support Analysis

Purpose: Analyze customer support tickets to identify trends and improvement opportunities

Operators Used:

  1. TextToSQL (get recent tickets)
  2. SQLExecution (retrieve data)
  3. PandasAi (initial analysis)
  4. TableToGraph (visualize volumes by category)
  5. OpenSearch (generate insights)
  6. CardGenerator (create summary dashboard)

Flow:

TextToSQL → SQLExecution → PandasAi → [TableToGraph, OpenSearch] → CardGenerator

Example 2: Document-Based Research

Purpose: Research a specific topic across a large document collection

Operators Used:

  1. RAGSubQuestions (break down research question)
  2. RAGSummarize (retrieve relevant information)
  3. TableToTextSummary (organize findings)
  4. Flow.Condition (check if more research needed)
  5. TextToGraph (visualize relationships)
  6. TextSummarize (create executive summary)

Flow:

RAGSubQuestions → RAGSummarize → TableToTextSummary → Flow.Condition
then_ → TextToGraph → TextSummarize
else_ → RAGSubQuestions (with refined questions)

Example 3: Automated Financial Reporting

Purpose: Generate periodic financial reports with analysis and visualizations

Operators Used:

  1. TextToSQL (extract financial data)
  2. SQLExecution (run queries)
  3. Code.Python.Execute (calculate financial metrics)
  4. TableToGraph (create financial charts)
  5. OpenSearch (generate narrative)
  6. ExcelToPPT (create presentation)

Flow:

TextToSQL → SQLExecution → Code.Python.Execute → [TableToGraph, OpenSearch] → ExcelToPPT

Next Steps

Now that you understand how to build minds in Quantum Minds, explore the following resources:

Overview | Operator Categories | Mind Templates | Integration Guide