NexusMind

NexusMind provides a Model Context Protocol (MCP) server that enables you to perform advanced scientific reasoning using a graph-based reasoning engine. It exposes an MCP endpoint for integration with compatible clients, supports structured graph queries, and delivers confidence-aware results with rich provenance across disciplines. This guide walks you through practical usage, installation steps, and essential configuration so you can deploy and use the MCP server effectively.

How to use

You connect your MCP-enabled client to the NexusMind MCP server to submit scientific queries and receive structured reasoning results. Start by ensuring the server is reachable at the configured MCP endpoint, then initiate a query with your desired scope and constraints. The server processes the request through its 8-stage reasoning pipeline, leverages graph-based relationships, and outputs a coherent answer along with supporting evidence and confidence assessments.

How to install

Prerequisites you need on your workstation before installing NexusMind MCP Server are Python 3.13+, Poetry for dependency management, and Docker for containerized deployment if you prefer a containerized setup.

Step by step local installation and run flow:

1. Clone the project repository to your working directory.
2. Install dependencies using Poetry.
3. Create and configure the application settings.
4. Start the server using the development command or a production-ready command for your environment.

Concrete commands to execute in your shell:

# Clone the repository
git clone https://example.org/nexusmind.git
cd nexusmind

# Install dependencies with Poetry
poetry install

# Activate the virtual environment (optional but recommended)
poetry shell

# Copy example settings and edit as needed
cp config/settings.example.yaml config/settings.yaml
vim config/settings.yaml

# Start the server in development mode
python src/asr_got_reimagined/main.py

# Or start with a live-reload web server
uvicorn asr_got_reimagined.main:app --reload --host 0.0.0.0 --port 8000

# The MCP endpoint will be available at /mcp

Additional sections

Configuration and deployment details are provided directly below to help you tailor the server to your environment. The MCP config for the NexusMind server is exposed as an HTTP endpoint that clients can reach to submit queries and retrieve results. The server supports a modular, extensible design with a FastAPI backend, Graph-of-Thoughts reasoning, and multi-dimensional confidence assessment.

Security and access controls should be considered for production deployments. Use environment variables and network policies to restrict who can access the MCP endpoint. Ensure sensitive configuration values are stored securely and not committed to version control.

If you plan to run the server in a container, you can use Docker Compose to orchestrate the API service, the reasoning engine, and any supporting data stores. The containerized setup simplifies deployment and scaling in both development and production environments.

Available tools

ASR-GoT Engine

Core graph-based reasoning engine that processes queries through an eight-stage pipeline to generate hypotheses, gather evidence, and produce final insights.

Graph Knowledge Representation

Uses graph structures with nodes, edges, and hyperedges to model relationships and support complex reasoning across domains.

MCP Protocol Integration

Exposes an MCP endpoint to integrate with Claude Desktop and other MCP clients for structured inquiry and response.

Confidence Vector Management

Maintains multi-dimensional confidence scores across empirical, theoretical, methodological, and consensus dimensions.

Evidence Gating and Pruning

Dynamically selects high-value evidence, prunes low-value elements, and consolidates related nodes to optimize graphs.

Subgraph Extraction

Identifies high-value subgraphs to focus analysis and extract targeted insights.

Composition and Citation

Synthesizes findings into a coherent narrative with node IDs, edge types, and citations.

Reflection and Quality Audit

Performs a comprehensive review of coverage, bias, and methodological rigor with final confidence assessment.