Visum Thinker MCP Server
Provides structured sequential thinking capabilities for AI assistants to break down complex problems into manageable steps, revise thoughts, and explore alternative reasoning paths.
README Documentation
Visum Thinker MCP Server
A Model Context Protocol (MCP) server that provides PTV Visum automation and sequential thinking capabilities for transportation planning and analysis.
šÆ Quick Start
New to Visum automation? ā Start here: QUICKSTART_PRT_WORKFLOW.md
AI Assistant integration? ā Read: CLAUDE_WORKFLOW_GUIDE.md
Full Documentation? ā See: DOCUMENTATION_INDEX.md
⨠New Features (2025-10-10)
š Interactive Visum Procedure Creation with Auto-Organization
Create and configure PTV Visum procedures with automatic organization in "Visum-BOT" group:
// 1. Create procedure ā Automatic group creation + delete operation
visum_create_procedure({procedureType: "PrT_Assignment"})
// ā Returns:
// - group_position: 577 (Visum-BOT group)
// - delete_position: 580 (Initialize Assignment)
// - actual_position: 581 (PrT Assignment) ā Use this!
// 2. List available demand segments (numbered 1-36)
visum_list_demand_segments()
// 3. Configure with user's choice
visum_configure_dsegset({
procedurePosition: 581, // Use actual_position!
segmentNumbers: "1-10" // or filterMode: "C", etc.
})
⨠Automatic Features:
- š¦ Group Organization: All MCP operations in "Visum-BOT" group
- šļø Auto-Delete: Initialize Assignment added before PrT/PuT assignments
- š Group Reuse: Subsequent calls add to existing group
- š Smart Positioning: Operations added at end of group
š Documentation: See VISUM_BOT_GROUP.md for complete details
**Key Features:**
- ā
**Automatic position detection** - No manual position tracking
- ā
**Numbered segment selection** - Easy "1-10" notation instead of long codes
- ā
**4 flexible input formats** - Numbers, mode filter, ALL keyword, or explicit codes
- ā
**Interactive workflow** - AI assistants guide users through configuration
- ā
**Complete validation** - Automatic verification of all configurations
**See [QUICKSTART_PRT_WORKFLOW.md](https://github.com/multiluca2020/visum-thinker-mcp-server/blob/main/QUICKSTART_PRT_WORKFLOW.md) for the 3-step workflow!**
## Features
### š Visum Transportation Planning
- **Procedure Management**: Create PrT/PuT Assignment, Demand Models, Matrix Calculations
- **Demand Segments**: List and configure demand segments with numbered selection
- **Interactive Configuration**: Guide users through DSEGSET setup with 4 input formats
- **Smart Position Tracking**: Automatically track actual procedure positions
- **Complete Validation**: Verify all configurations before execution
### š§ Sequential Thinking
- **Step-by-step reasoning**: Break down complex problems into sequential thoughts
- **Dynamic revision**: Revise and refine thoughts as understanding deepens
- **Branching logic**: Branch into alternative reasoning paths
- **Adaptive planning**: Adjust the total number of thoughts dynamically
- **State management**: Maintain thinking context across multiple tool calls
- **Progress tracking**: Monitor completion status and thought progression
### š PDF Analysis
- **PDF Loading**: Load and analyze PDF documents for problem-solving context
- **Content Search**: Find relevant sections in PDFs based on queries and search terms
- **Persistent Storage**: Auto-save state to disk, survive server restarts
- **Knowledge Transfer**: Export/import thinking sessions between servers
## Installation
### Quick Installation
```bash
# Option 1: Install from NPM
npm install -g visum-thinker-mcp-server
# Option 2: Use with npx (no installation)
npx visum-thinker-mcp-server
# Option 3: Clone from GitHub
git clone https://github.com/yourusername/visum-thinker-mcp-server.git
cd visum-thinker-mcp-server
npm install && npm run build
See INSTALLATION.md for detailed setup instructions.
Prerequisites
- Node.js 16 or higher
- npm or yarn
Usage
With Claude Desktop
Add to your Claude Desktop configuration (claude_desktop_config.json):
{
"mcpServers": {
"visum-thinker": {
"command": "node",
"args": ["/absolute/path/to/sequential_thinking/build/index.js"]
}
}
}
With VS Code
The project includes a .vscode/mcp.json configuration file for VS Code MCP integration.
Direct Usage
npm run dev
Tools
sequential_thinking
Main tool for step-by-step reasoning process.
Parameters:
thought(string): The current thinking stepnextThoughtNeeded(boolean): Whether another thought step is neededthoughtNumber(integer): Current thought numbertotalThoughts(integer): Estimated total thoughts neededisRevision(boolean, optional): Whether this revises previous thinkingrevisesThought(integer, optional): Which thought is being reconsideredbranchFromThought(integer, optional): Branching point thought numberbranchId(string, optional): Branch identifierneedsMoreThoughts(boolean, optional): If more thoughts are needed
load_pdf
Load a PDF file to provide context for analysis.
Parameters:
filePath(string): Absolute path to the PDF file
analyze_pdf_section
Analyze specific sections of the loaded PDF.
Parameters:
query(string): What to look for or analyze in the PDFstartPage(integer, optional): Starting page number (1-based)endPage(integer, optional): Ending page number (1-based)searchTerms(array of strings, optional): Specific terms to search for
reset_thinking
Clears the current thinking state to start fresh.
get_thinking_summary
Returns a summary of the current thinking session including PDF context if loaded.
export_knowledge
Export the current thinking state and PDF knowledge to a file.
Parameters:
exportPath(string): Absolute path where to save the exported knowledge file
import_knowledge
Import thinking state and PDF knowledge from an exported file.
Parameters:
importPath(string): Absolute path to the exported knowledge file to import
š Visum Transportation Planning Tools
The server includes comprehensive PTV Visum integration with intelligent path learning:
check_visum: Check Visum availability and learn custom installation pathsload_visum_model: Load transportation models (.ver files)run_visum_calculation: Execute transportation calculations and analysesget_network_statistics: Analyze network topology and characteristicsanalyze_visum_matrices: Examine demand and flow matricesexport_visum_results: Export analysis results to various formats
Key Features:
- š§ Smart Path Learning: Automatically remembers custom Visum installation paths
- š Zero Setup: Works seamlessly after initial path discovery
- šÆ Demo Mode: Full testing capability without Visum installation
- š Complete Analysis: All major transportation planning workflows supported
See VISUM-PATH-LEARNING.md for detailed information about the intelligent path learning system.
š¤ GitHub Copilot Integration
The Sequential Thinking MCP Server includes comprehensive GitHub Copilot integration for enhanced AI-assisted development:
š Quick Start with Copilot
- Server Status: Ensure MCP server is running (
npm run dev) - Open Copilot Chat:
Ctrl+Shift+Iin VS Code - Test Integration: Ask
@copilot List available MCP tools - Start Solving:
@copilot Use sequential thinking to solve [your problem]
šÆ Copilot Capabilities
- š§ Sequential Thinking: AI-guided step-by-step problem solving
- š PDF Analysis: Intelligent document processing and analysis
- š Transportation Planning: Expert Visum integration and workflow automation
- š§ Smart Configuration: Automatic Visum path learning and persistence
- š” Context-Aware Suggestions: Code completion with domain knowledge
š¬ Example Copilot Interactions
@copilot Can you use sequential thinking to analyze this transportation network problem?
@copilot Check if Visum is available and help me load a network model
@copilot Use the PDF analysis tools to extract data from this traffic report
@copilot Create a complete workflow for transportation demand analysis
See COPILOT-INTEGRATION.md for comprehensive setup and usage guide.
Development
Project Structure
sequential_thinking/
āāā src/
ā āāā index.ts # Main server implementation
āāā build/ # Compiled JavaScript output
āāā .vscode/
ā āāā mcp.json # VS Code MCP configuration
āāā .github/
ā āāā copilot-instructions.md
āāā package.json
āāā tsconfig.json
āāā README.md
Scripts
npm run build: Compile TypeScript to JavaScriptnpm run dev: Build and run the servernpm test: Run tests (placeholder)
Debugging
The server logs to stderr for compatibility with STDIO transport. Use VS Code's debugging features or add console.error statements for debugging.
Architecture
The server maintains a global thinking state that tracks:
- All thoughts in the current session
- Current progress and estimated completion
- Revision and branching relationships
- Session completion status
Each tool call updates this state and provides formatted responses that help users follow the thinking process.
License
MIT