Knowledge Graphs for LLM RAG Workflows
Powerful GraphRAG API that helps your LLM an AI agents understand context, discover gaps, and generate unique insights.
Watch an Introduction
Watch an Introduction
InfraNodus API: Plug-and-Play GraphRAG for AI Agents
InfraNodus represents any text as a knowledge graph. It can then be used to improve LLM prompts, optimize knowledge bases, and improve the quality of retrieval augmented generation (RAG) AI workflows.
Sign Up to Get an API Key
How InfraNodus Can Help You
You can use our visual interface to get an overview of your AI knowledge base, retrieve the main topics and concepts, see what's missing. Get a GraphRAG summary to augment your system LLM prompts.
Automatically generated knowledge graph can extract ontology and relations from any text and use them to augment your RAG workflows with high-quality relational information.
Find the content gaps: find the topical clusters and ideas that are not yet linked, use them to optimize your knowledge base or to feed them to LLM agents to improve the quality of their responses.
Use our API to generate research questions that can be used in your LLM workflows. Feed these questions to AI agents to conduct better research.
Import Content from Multiple Sources
You can send your text with a simple API call or use our visual interface to import content from multiple sources and formats automatically, no need for a scraper anymore. Here are some of the most popular ones:
Use It to Build AI Chatbots, AI Research Agents, and to Augment User Prompts
InfraNodus GraphRAG API can be used for generating research questions, optimizing knowlege bases, improving user prompts, creating the tools that boost the intelligence of your AI agents.
-
Improve RAG Responses
Augment user prompts with contextual data.
Watch the Video -
Plug-and-Play GraphRAG
Improve the quality of retrieval.
Learn More -
Knowledge Optimization
Improve the quality of your LLM context.
Case Study -
Content Gap AI Agent Tools
Detect content gaps and generate questions.
Case Study
Use Case: Building an AI-powered chatbot for a customer support portal
Problem: Your standard AI chatbot won't answer general questions well
Standard AI RAG workflows are not suitable for answering general questions. They will search for the content similar to the user query, but they don't understand the meaning of the questions like "What is it for"?
Solution: Use the InfraNodus GraphRAG API to improve LLM RAG responses
The InfraNodus API will extract the main topics and keywords from your knowledge base (`graphSummary` variable in the `graphAndAdvice` endpoint), which you can then use to augment user prompts and make them more pertinent to the content they query.
Augment User Prompts Case Study Try It Now
Use Case: Teach your LLMs to reason using the GraphRAG / HyperRAG functionality
Problem: Standard LLM responses do not take relations into account
Most LLM worksflows treat text as separate chunks of information and miss out the relations within the text.
Solution: Use InfraNodus GraphRAG to add relational context to your LLM responses
InfraNodus will convert user prompts into graphs and overlay them on top of the knowledge base graph. It will then follow the graph's edges to retrieve important relational context and add it to the LLM context window to improve the response quality.
GraphRAG Case Study
Try It Now
Use Case: Optimize your knowledge base for LLM context
Problem: Your knowedge base might be incomplete or biased
If the knowledge base that you use to provide the context to your LLM lacks certain topics or is too focused on a specific domain, the quality of responses will suffer.
Solution: Use InfraNodus to optimize your knowledge base with network analysis and visualization
InfraNodus will represent your knowledge base as a knowledge graph and show the main topics, concepts, and relations within. You can use this information to get an overview of the content and identify what's missing. Structural analysis will help identify bias. Content gap detection will reveal the blind spots. Target those to improve the quality of your LLM context.
Knowledge Base Optimization Case Study Try It Now
Use Case: Add creative thinking tools to your AI agents
Problem: Standard LLM agents are not good in detecting content gaps
As a result, they provide generic responses and miss out on important research and marketing opportunities.
Solution: Integrate InfraNodus content gap detection tools into your AI agent workflows
InfraNodus tools can be used by your AI agents to detect content gaps and steer your model towards bridging those gaps to generat new ideas. InfraNodus can also generate research questions that can be used as prompts in your AI workflows.
Crew AI Content Gap Case Study Try It Out
Examples of Knowledge Graphs
These are examples of the knowledge graphs. The visualizations can be used to get an overview and optimize your knowledge base. The graphs can also be embedded into your LLM workflow using the InfraNodus MCP server. This makes it possilbe to add context, retrieve and save memories in your LLM conversations, and use as as portable GraphRAG knowledge base.
Start Using InfraNodus
Sign up for an account now, so you can start using InfraNodus to augment your LLM workflows.
Knowledge Graph for LLM RAG: How InfraNodus Compares
There are several approaches to using knowledge graphs for LLM RAG workflows. Most solutions require you to set up and manage a graph database, write custom extraction pipelines, or build your own retrieval logic. InfraNodus takes a different approach: it provides a portable, plug-and-play GraphRAG API that works out of the box with any LLM framework.
Other Knowledge Graph RAG Approaches
- Microsoft GraphRAG requires setting up a local Python environment, configuring LLM API keys, and running indexing pipelines before you can query your data.
- Neo4j Graph Builder requires deploying and managing a Neo4j database instance and writing Cypher queries for retrieval.
- LangChain Graph Transformer provides graph extraction utilities but requires you to build the full retrieval and summarization pipeline yourself.
InfraNodus Knowledge Graph API for LLM RAG
- No graph database needed — send text via API, get a knowledge graph with topics, relations, and gaps in seconds.
- Built-in content gap detection — automatically identifies what is missing in your knowledge base, so your AI agents can focus on bridging those gaps.
- Visual observability — an interactive graph interface lets you inspect the knowledge base structure, unlike black-box pipelines.
- Works with any LLM framework — Dify, Crew AI, n8n, custom Python/Node.js agents, or direct API calls.
How Does InfraNodus Work?
InfraNodus visualizes any text as a knowledge graph. Drawing an analogy from social sciences, words are represented as nodes, while their co-occurrences create the connections between them. This network representation enables the use of advanced graph theory algorithms to identify clusters of related ideas, highlight the most influential concepts and hidden intermediaries, and reveal gaps within the discourse. This approach can be used to enhance your perspective on any discourse and explore nuances of meaning that might otherwise remain hidden.
Additionally, AI-driven algorithms are employed to generate insights and new ideas for any discourse based on the topical clusters and gaps discovered.
To learn more about the algorithms used in InfraNodus and for citations, please, refer to this peer-reviewed paper:
Paranyushkin, D (2019). InfraNodus: Generating Insight Using Text Network Analysis, Proceedings of WWW'19 The Web Conference, www.infranodus.com (ACM library, PDF).
Ecological Thinking and Cognitive Variability
Gregory Bateson coined a beautiful term: "ecology of the mind". What is a mind that is ecological? It has the ability to have an overview, but it can also zoom in on any idea. It embraces diversity, but it can also obsess over one thing when needed. It can discover the obvious, but it can also reveal the hidden and ponder the gaps that have yet to be bridged. Focused, yet adaptive. Rational, yet poetic.
InfraNodus is a tool that is developed to help you think this way. It is made to promote ecological dynamics and diversity on a cognitive level. Too often, people get stuck in certain ideas or thought patterns. InfraNodus can be used as a "mind antivirus" against obsessive loops — biased ideas, mundane patterns, totalitarian thinking, propaganda, and narrow-mindedness — proposing a pan~archic form of thinking that spans a range of states and modalities.
Sign Up for a Free 2-Week Trial
Discover the power of ecological thinking and cognitive variability using knowledge graphs and AI-driven insights.
Pricing Options
Your support is what makes it possible for us to develop this tool. All quotas are per each text. There is no limitation on the number of texts you can process every month. Prices do not include VAT for EU private customers.
Lock in current prices for life and save up to 35% (€100-€200) on our annual plans.
Advanced Account
saving €204 or 35%
- 14-day free trial
- everything on the Basic +
- API access
- Dedicated support
- 2 Mb per upload
- 5 Mb per PDF upload
- Extended data import quotas
- 100 GPT-4 credits / hour
- Live graph updates (max 5)
- Commercial use
Basic Account
saving €84 or 35%
- 14-day free trial
- Community support
- Full graph analytics
- Chrome / Firefox extension
- Obsidian graph view plugin
- Import data from web sources
- Max 40 imports per source
- 300 Kb per file upload
- 1Mb per PDF upload
- 40 GPT-4 AI credits / hour
- Personal / Academic use
Premium Account
- 14-day free trial
- everything on Advanced +
- API integration support
- Training: 1 hour / month
- 10 Mb per upload
- 50 Mb per PDF upload
- Max import quotas
- 500 GPT-4 credits / hour
- Live graph updates (max 20)
- Fast-track required features
InfraNodus vs. Other Tools
Traditional LLM tools and AI chatbots are limited by their inability to understand the underlying structure and relationships in your data. InfraNodus provides a unique approach to text analysis that reveals the hidden patterns and insights in your data.
other tools
- Standard RAG workflows that treat your text as disjointed data chunks
- Generic, predictable responses
- Cannot deal with general queries
- Black-box approach hides reasoning
- No graph, no deeper understanding of context
- Take days to set up
InfraNodus
- GraphRAG that understands relations in your data and provides rich context
- Innovative responses that focus on content gaps
- Has a holistic view of your data
- Interactive graph shows you what's happening
- GraphRAG finds gaps in your knowledge
- Take minutes to set up
Testimonials from Our Users
InfraNodus API is used by researchers and consultants to augment their LLM workflows. Here are some of the testimonials from our users:
Frequently Asked Questions about Knowledge Graphs and LLM RAG
How can knowledge graphs help improve LLM RAG workflows?
Traditional RAG (retrieval-augmented generation) workflows treat your text as disjointed data chunks and provide generic, predictable responses. Knowledge graphs, on the other hand, provide a holistic view of your data by representing it as a graph of concepts and relationships. This allows the LLM to understand the context of your data and generate more accurate, nuanced responses.
What is GraphRAG and how does it improve LLM responses?
GraphRAG (Graph-based Retrieval Augmented Generation) enhances standard RAG by representing your data as a knowledge graph rather than isolated text chunks. This allows the LLM to understand relationships between concepts, retrieve richer context, and generate more accurate, nuanced responses. InfraNodus provides a plug-and-play GraphRAG API that extracts topics, relations, and content gaps from any text, improving the quality of LLM outputs.
How do you use a knowledge graph for RAG?
To use a knowledge graph for RAG, you send your text data to a knowledge graph engine (like InfraNodus) via API. The engine extracts entities and their relationships, forming a graph structure. When a user query comes in, the graph is traversed to retrieve not just similar text chunks, but also the relational context between concepts. This enriched context is then passed to the LLM alongside the user query, resulting in more informed and accurate responses.
What is the difference between GraphRAG and traditional RAG?
Traditional RAG splits documents into chunks and retrieves them using vector similarity search. It works well for specific queries but struggles with general questions that require understanding the broader context. GraphRAG adds a knowledge graph layer that captures the relationships between concepts, enabling holistic understanding of the data. This means GraphRAG can answer questions like "What is this about?" or "What is missing?" that traditional RAG cannot handle effectively.
How can AI agents use knowledge graphs?
AI agents can use knowledge graphs to gain a structured understanding of their domain, detect content gaps and blind spots, and generate targeted research questions. With InfraNodus, AI agents can call the API to retrieve topical clusters, discover what information is missing, and generate questions that bridge those gaps. This makes agents more autonomous and capable of producing original insights rather than generic responses.