# InfraNodus MCP Server

> MCP server for knowledge graph generation, text network analysis, content gap detection, and SEO optimization powered by InfraNodus API and graph theory algorithms.

## Documentation

- [README](https://github.com/infranodus/mcp-server-infranodus/blob/main/README.md): Installation, setup, and feature overview
- [Tool Examples](https://github.com/infranodus/mcp-server-infranodus/blob/main/EXAMPLES.md): Detailed examples for every tool with request/response formats
- [Full LLM Documentation](https://github.com/infranodus/mcp-server-infranodus/blob/main/llms-full.txt): Comprehensive tool documentation optimized for LLMs
- [API Access](https://infranodus.com/api-access): Get your InfraNodus API key
- [InfraNodus MCP](https://infranodus.com/mcp): MCP integration overview

## Tools

### Analysis

- generate_knowledge_graph: Convert text into a knowledge graph with topics, concepts, relations, and structural analysis
- analyze_text: Analyze text, URL, or YouTube transcript for topics, clusters, and structural insights
- analyze_existing_graph_by_name: Retrieve and analyze existing graphs from your InfraNodus account
- generate_topical_clusters: Extract compact topical clusters from text
- generate_content_gaps: Identify missing connections between topic clusters
- generate_contextual_hint: Generate structural summary for RAG/LLM augmentation

### Idea Generation

- generate_research_questions: Generate research questions that bridge content gaps
- generate_research_ideas: Generate innovative ideas based on content gaps
- optimize_text_structure: Analyze bias/coherence and suggest development directions
- develop_text_tool: Comprehensive text development combining multiple analysis tools
- develop_conceptual_bridges: Connect discourse to broader context via gateway concepts
- develop_latent_topics: Find and develop underdeveloped topics
- generate_responses_from_graph: Generate AI responses based on existing knowledge graph

### Comparison

- overlap_between_texts: Find common topics across multiple texts
- difference_between_texts: Find what's missing in first text vs others
- merged_graph_from_texts: Merge multiple sources into one graph with combined analysis

### SEO & Search

- generate_seo_report: Full SEO analysis comparing text with search results and queries
- analyze_google_search_results: Graph of Google search results for queries
- analyze_related_search_queries: Graph of related search queries (search intent)
- search_queries_vs_search_results: Find what people search for but don't find

### Memory & Storage

- create_knowledge_graph: Save text as a knowledge graph in InfraNodus
- memory_add_relations: Save statements to memory: a named graph (written immediately), or with scope the user's kept topic memory (memory-<topic>) or personal memory (personal-<user>), dry run until confirm: true
- memory_get_relations: Recall from memory by entity, across graphs or one graph; with scope from the kept topic or personal memory (entity, prompt, or overview)
- enable_memory: Enable topic or personal memory (only on the user's request)
- enable_project_learnings: Enable the per-project learnings graph (learn-<project>), only on the user's request
- add_project_learnings: Save what was learned about operating in a project; dry run until confirm: true
- get_project_learnings: Retrieve a project's learnings for a task (prompt), an entity, or as an overview
- retrieve_from_knowledge_base: GraphRAG retrieval from existing knowledge base

### Utility

- list_graphs: List all graphs with filtering by name, type, date, language
- search: Search through existing InfraNodus graphs
- fetch: Fetch specific search results from a graph
- delete_statements: Delete statements from your own graph by one filter (categories, statements, query, before/after, deleteAll, statementIds); dry run unless confirm: true; irreversible
- update_statements: Edit statements of your own graph in place (content, categories, timestamp) keeping id and position — specific statements by exact text or id (edits), or one selector with set/replace for bulk relabelling and find-and-replace; dry run showing before → after unless confirm: true; irreversible

## Optional

- [npm Package](https://www.npmjs.com/package/infranodus-mcp-server): NPM package for local installation
- [Smithery](https://smithery.ai/server/@infranodus/mcp-server-infranodus): Alternative installation via Smithery
- [InfraNodus Research](https://noduslabs.com/research/): Graph theory concepts and methodology
