Zeno Mind

MCP Protocol Integration

Learn how ZenoMind Desktop leverages Model Context Protocol for tool integration.

What is MCP?

Model Context Protocol (MCP) is an open standard that enables AI assistants to securely connect with various tools and data sources. ZenoMind Desktop uses MCP to integrate with search, file system, and command line tools.

Available MCP Tools

ZenoMind Desktop supports these MCP tools for enhanced functionality:

Search Tools

  • Web Search: Search the internet for information
  • Local Search: Search files and content on your system
  • Research Integration: Gather information from multiple sources

File System Tools

  • File Operations: Read, write, and manage local files
  • Directory Navigation: Browse folder structures
  • File Content Search: Search within file contents
  • Document Processing: Handle various document formats

Command Line Integration

  • Terminal Access: Execute command line operations
  • Script Execution: Run automation scripts
  • System Operations: Perform system-level tasks
  • Development Tools: Access development utilities

Setting Up MCP Tools

Initial Configuration

When you first launch ZenoMind Desktop, configure your tools:

  1. Model Configuration: Set up your AI model provider (OpenAI, Anthropic, etc.)
  2. Search Configuration: Configure web search providers
  3. Tool Permissions: Set access levels for file system and command line tools

Model Provider Setup

OpenAI Configuration

  • API Key
  • Model selection
  • Usage preferences

Anthropic Configuration

  • API Key
  • Model selection
  • Claude-specific settings

Azure OpenAI (if supported)

  • API Version
  • Deployment Name
  • Endpoint URL

Search Provider Setup

Configure your preferred search providers:

  • Search engine selection
  • API key configuration
  • Search scope settings

Using MCP Tools

Automatic Tool Selection

ZenoMind Desktop automatically selects appropriate tools based on your requests:

"Search for information about renewable energy trends"
→ Uses web search tools automatically

"Find all Python files in my project folder"
→ Uses file system tools to search locally

Tool Workflows

Research Workflow

  1. Web search for information
  2. Analyze and summarize findings
  3. Save results to local files
  4. Generate comprehensive reports

File Management Workflow

  1. Search for specific files
  2. Read and analyze content
  3. Edit or organize files
  4. Execute related commands

Development Workflow

  1. Search codebase for issues
  2. Edit files as needed
  3. Run tests via command line
  4. Manage project files

Real-time Artifacts

View tool operations in real-time:

  • Search Results: See web search results as they're found
  • File Operations: Monitor file reading and writing
  • Command Execution: Watch command line operations
  • Document Processing: View document analysis progress

Human-in-the-Loop Interaction

Interactive Control

During tool execution, you can:

  • Interrupt Operations: Stop ongoing processes
  • Provide Guidance: Redirect the AI's approach
  • Insert Feedback: Add your input during execution
  • Modify Direction: Change the task focus mid-process

Collaboration Features

  • Real-time Input: Provide feedback while the AI works
  • Process Steering: Guide the AI's decision-making
  • Quality Control: Review and approve actions
  • Learning Integration: Help the AI understand your preferences

Thread Sharing

Local Sharing

Share your conversations as local HTML files:

  1. Click the share button in the interface
  2. Select "Local HTML" mode
  3. Save the generated HTML file
  4. Share the file with others

Remote Sharing

Share via remote URL (if configured):

  1. Set up remote server configuration
  2. Click share and select "Remote URL"
  3. The system uploads your thread to the server
  4. Share the generated URL with others

Best Practices

Security Guidelines

  1. Permission Management: Only grant necessary tool permissions
  2. Data Protection: Be cautious with sensitive file access
  3. Command Safety: Review potentially dangerous commands
  4. Access Control: Regularly review tool access settings

Performance Tips

  1. Efficient Queries: Be specific in your requests
  2. Tool Selection: Understand which tools work best for specific tasks
  3. Resource Management: Monitor system resource usage
  4. Workflow Optimization: Learn effective tool combinations

Troubleshooting

Common Issues

  • Tool connection failures
  • Permission denied errors
  • API key configuration issues
  • Network connectivity problems

Debugging Steps

  1. Check tool configuration in settings
  2. Verify API keys and permissions
  3. Test tools individually
  4. Review error messages for specific issues

Limitations and Considerations

Current Limitations

  • macOS support only (other platforms coming soon)
  • Internet connection required for web search and AI models
  • Some tools may require specific permissions
  • Performance depends on system resources

Technical Preview Notice

ZenoMind Desktop is currently in preview. Features may change and some functionality may be unstable. We welcome feedback for improvements.

Ready to explore more features? Check out our User Interface Guide for detailed usage instructions.

Features