LogicMap MCP Server
# LogicMap MCP Server
An MCP server implementation that provides semantic knowledge graph tools for understanding project structure, business logic, and code relationships.
## Features
- **Semantic Knowledge Graph**: Record business logic and conceptual relationships, not just code structure
- **Multi-Project Management**: Manage multiple project knowledge graphs simultaneously
- **Rich Node Types**: Support for concepts, flows, modules, components, data entities, and external dependencies
- **Relationship Modeling**: Express various logical relationships (calls, depends, implements, flows_to, uses_data, triggers, extends)
- **Powerful Query Tools**: Search nodes, query relationships, and explore graph structure
- **Language Agnostic**: No dependency on specific language parsers, works with any codebase
- **Progressive Construction**: Build knowledge graphs incrementally, starting with core flows
## Tools
### Project Management
- **createProject**: Create a new project with metadata
- **updateProject**: Update project information
- **deleteProject**: Remove a project
- **listProjects**: List all available projects
### Node Management
- **addNode**: Add a semantic node to the knowledge graph
- Types: `concept`, `flow`, `module`, `component`, `data`, `external`
- Supports tags, file associations, and custom metadata
- **updateNode**: Update existing node information
- **removeNode**: Remove a node and optionally cascade delete related edges
### Relationship Management
- **linkNodes**: Create relationships between nodes
- Types: `calls`, `depends`, `implements`, `flows_to`, `uses_data`, `triggers`, `extends`
- Supports relationship strength and descriptions
- **unlinkNodes**: Remove relationships between nodes
### Query Tools
- **queryNode**: Get detailed node information with relationships
- Supports neighbor traversal with configurable depth
- **searchNodes**: Search nodes by criteria
- Filter by type, tags, files, or text content
- Configurable result limits
## Usage
LogicMap is designed for:
- Understanding complex codebases and business logic
- Documenting system architecture and data flows
- Impact analysis before making changes
- Onboarding new team members
- AI-assisted code comprehension
- Cross-team knowledge sharing
## Configuration
### Usage with Claude Desktop
Add this to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"logicmap": {
"command": "node",
"args": [
"/absolute/path/to/logicmap-mcp-server/dist/index.js"
]
}
}
}
```
### Usage with Kiro (compatible with other vscode based IDE)
Add the configuration to your MCP configuration file. Open the Command Palette (`Ctrl + Shift + P` or `Cmd + Shift + P`) and run `MCP: Open User Configuration`. This will open your user `mcp.json` file where you can add the server configuration.
**User Configuration (Recommended)**:
```json
{
"mcpServers": {
"logicmap": {
"command": "node",
"args": [
"/absolute/path/to/logicmap-mcp-server/dist/index.js"
]
}
}
}
```
**Workspace Configuration**:
Alternatively, create `.kiro/settings/mcp.json` in your workspace:
```json
{
"mcpServers": {
"logicmap": {
"command": "node",
"args": [
"/absolute/path/to/logicmap-mcp-server/dist/index.js"
]
}
}
}
```
> For more details about MCP configuration, see the [Kiro MCP documentation](https://kiro.dev/docs/mcp).
### Configuration Example
See `mcp-config-example.json` for a complete configuration example.
## Installation
### From Source
```bash
# Clone the repository
git clone https://github.com/yourusername/logicmap-mcp-server.git
cd logicmap-mcp-server
# Install dependencies
npm install
# Build the project
npm run build
```
### Verify Installation
After building, test the server:
```bash
# Run tests
npm test
# Check build output
ls dist/
```
## Development
```bash
# Development mode (watch for changes)
npm run dev
# Run tests
npm test
# Run tests with coverage
npm run test:coverage
# Lint code
npm run lint
# Format code
npm run format
# Clean build output
npm run clean
```
## Example Usage
### Creating a Knowledge Graph
```typescript
// 1. Create a project
await createProject({
name: "My Web App",
description: "E-commerce platform",
workspacePath: "/path/to/project"
});
// 2. Add semantic nodes
await addNode({
id: "user-auth-flow",
type: "flow",
name: "User Authentication Flow",
description: "Handles user login, token validation, and permission checks",
files: ["src/auth/controller.ts", "src/auth/service.ts"],
tags: ["core", "security"]
});
await addNode({
id: "payment-service",
type: "module",
name: "Payment Service",
description: "Integrates with Stripe and PayPal for payment processing",
files: ["src/payment/processor.ts"],
tags: ["core", "third-party"]
});
// 3. Create relationships
await linkNodes({
from: "user-auth-flow",
to: "payment-service",
type: "depends",
description: "Payment requires authenticated user session"
});
// 4. Query the graph
const result = await queryNode({
id: "user-auth-flow",
includeNeighbors: true,
depth: 2
});
// 5. Search nodes
const searchResults = await searchNodes({
tags: ["core"],
types: ["flow", "module"],
limit: 10
});
```
## Project Structure
```
logicmap-mcp-server/
├── src/
│ ├── index.ts # Entry point
│ ├── types/ # TypeScript type definitions
│ ├── schemas/ # JSON Schema validation
│ ├── services/ # Core business logic
│ │ ├── NodeService.ts # Node CRUD operations
│ │ ├── EdgeService.ts # Relationship management
│ │ └── QueryService.ts # Graph queries
│ ├── storage/ # Data persistence
│ │ └── StorageManager.ts # File-based storage with locking
│ └── mcp/ # MCP protocol implementation
│ └── MCPServerHandler.ts
├── tests/ # Test files
├── dist/ # Build output
└── package.json
```
## Data Storage
LogicMap stores project data in `~/.logicmap/projects/` as JSON files:
```
~/.logicmap/
└── projects/
└── {project-id}/
└── graph.json # Complete knowledge graph
```
Each graph file contains:
- Project metadata
- Node dictionary (for O(1) lookups)
- Edge list
- Timestamps
## Tech Stack
- **Runtime**: Node.js >= 18.0.0
- **Language**: TypeScript 5.3+
- **Protocol**: Model Context Protocol (MCP)
- **Validation**: Ajv (JSON Schema)
- **Storage**: JSON files with proper-lockfile
- **Testing**: Vitest + fast-check
## Building
```bash
# Build TypeScript to JavaScript
npm run build
# Output will be in dist/ directory
```
## Contributing
Contributions are welcome! Please see [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
## License
This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the [LICENSE](LICENSE) file in the project repository.
## Links
- [Model Context Protocol Documentation](https://modelcontextprotocol.io)
- [MCP TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk)
- [Issue Tracker](https://github.com/Sim-xia/Logic-map-MCP/issues)
TDQS
Scored across 11 tools
Each tool targets a distinct resource and action: project lifecycle tools (create, update, delete, list) are clearly separate from knowledge graph operations (add, update, remove, link, unlink, query, search). No two tools appear to overlap in purpose.
All tool names follow a consistent verbNoun camelCase pattern: createProject, updateProject, deleteProject, listProjects, addNode, updateNode, removeNode, linkNodes, unlinkNodes, queryNode, searchNodes. The pattern is uniform and predictable.
With 11 tools, the set is well-scoped for a server managing both projects and a knowledge graph. Each tool serves a distinct purpose and the count feels appropriate, not bloated or sparse.
The surface provides full CRUD coverage for projects (create, update, delete, list) and comprehensive graph operations for nodes (add, update, remove, link, unlink, query, search). The lifecycle is complete with no obvious dead ends or missing core functionality.