12 specialized MCP servers providing AI agents with structured access to research databases, infrastructure tools, and workspace integrations.
Each server exposes domain-specific tools and resources via the Model Context Protocol.
Docker management, system monitoring, and infrastructure automation. Query container status, manage networks, inspect services, and execute operational tasks.
infra-mcp.homeInterface to Google NotebookLM for AI-powered document analysis, source management, and audio overview generation. Manages notebooks and their sources programmatically.
notebooklm-mcpSearch, retrieve, and analyze scientific papers from arXiv, Semantic Scholar, and other academic databases. Supports citation graphs and full-text extraction.
sci-papers-mcpGoogle Workspace integration for Gmail, Drive, Calendar, and Docs. Read emails, manage files, create events, and search across workspace content.
gws-mcpQuery Grafana dashboards, data sources, and alerts. Retrieve metrics, create annotations, and interact with monitoring data through structured tools.
grafana.homeBrowse and search HuggingFace models, datasets, and spaces. Download model cards, check benchmarks, and discover AI/ML resources.
huggingface-mcpSearch PubMed and PMC for biomedical literature. Retrieve abstracts, full-text articles, citation data, and MeSH term metadata.
pubmed-mcpGenerate Mermaid diagrams programmatically. Create flowcharts, sequence diagrams, Gantt charts, entity-relationship diagrams, and architecture views.
mermaid-mcpDirect database access for MariaDB/MySQL. Execute queries, inspect schemas, manage tables, and perform database administration through MCP tools.
mysql-mcpAI-powered transcript correction and cleanup. Fix speech recognition errors, normalize formatting, and improve readability of automated transcriptions.
transcript-correction-mcpSmart home integration via Home Assistant. Control devices, query entity states, trigger automations, and access sensor data through MCP tools.
ha-mcpInterface to the LibroSynth AI research platform. Trigger cross-domain synthesis, causal analysis, and knowledge fusion workflows via the orchestrator API.
librosynth-mcpAll MCP servers are aggregated behind a unified proxy at mcp.home. The proxy handles routing, authentication (via X-API-Key header), and health monitoring for all backends. Three backend groups: docker, ollama, and database.
The Model Context Protocol standardizes how AI agents interact with external tools and data.
Functions that agents can invoke — search, query, create, update. Each server exposes domain-specific tool definitions.
Structured data that agents can read — documents, schemas, configs. Exposed as URI-addressable content.
Pre-built prompt templates for common workflows. Agents select and parameterize prompts for specialized tasks.
Stdio, SSE, or Streamable HTTP. The n8n MCP server uses Streamable HTTP via supergateway for Claude Code integration.