Free
€0
- Try it on your own text
- Files and PDFs up to 1 MB
- 3 graphs and 3 AI chats
- No card required
Works inside ChatGPT and Claude
Discover what everyone else is missing
InfraNodus turns any text into a knowledge graph of ideas. You see the main topics, how they connect, and the gaps between them, where new ideas come from.
Are you one of them?
Hover a word or a dot
- Use case
- A knowledge graph your agents can reason with Connect InfraNodus to Claude, Cursor or your own agent over MCP. Graph what it knows about your codebase, docs or past sessions, and see what is missing before it acts.
- Found
- The gap: tool calls never reach memory: nothing a tool returns is stored or retrieved later.
- Next
- Ask Claude to close the gap, with InfraNodus connected over MCP.
- Use case
- Content gaps in search results and AI answers Map what Google ranks and what ChatGPT and Claude answer about a CRM for a small team, compare it with what people search for, and write what nobody covers yet.
- Found
- The gap: pricing and integrations never meet. No answer says what the integrations cost, and nobody ranks for it yet.
- Next
- Write that article, outlined from the gap.
- Use case
- The structure of your vault, and what it is missing Upload your Markdown notes, see the clusters and the notes that never link, develop the missing idea with AI, then bring it back with the Obsidian plugin.
- Found
- The gap: your notes on habits and on writing never link, though both are about showing up every day.
- Next
- Develop the bridge note with AI, then save it to your vault with the plugin.
- Use case
- Gaps in a literature review or 40 interviews Upload papers, transcripts or survey answers, explore their themes and how they connect, from price to support, and keep the best ideas in Project Notes as your framework grows.
- Found
- The gap: nobody connects onboarding to the cost of inviting clients.
- Next
- Save it as a finding in Project Notes and ask the next question.
Watch a text become a knowledge graph
A few paragraphs about AI agents, analysed the way InfraNodus analyses yours. Click through it at your own pace.
- 1Words become nodes
- 2Topics emerge as clusters
- 3The gap becomes a question
- 4Create unique insight
Source material · AI agentsAn example to explore
Starts automatically
A new perspective
on the same text.
Research question
Dots: words · Lines: used togetherDrag to rotate
Start with a few sentences. Watch their ideas take shape.
See what these ideas say together.
Topic summary
Select another topic to update this summary.
Now try it on your own notes
- notes
vault
- papers.pdf
- draft.md
- interviews.txt
- serp.csv
What our users say
“I keep wondering how many weeks or even months I would spend trying to collect all this knowledge by myself.”
“Both features were priceless for developing effective cross-constituency system design.”
“Its utility in SEO and market research, where it identifies gaps between informational demand and supply, is especially notable.”
Use it inside your AI
Claude, ChatGPT, Cursor or any MCP client gets graph tools: topics, gaps, research questions and your saved graphs.
Asked in Claude Code, with the infranodus MCP server connected: “Make a knowledge graph of my text files in the current folder and find what’s missing”
Tools used: infranodus generate_knowledge_graph, infranodus generate_content_gaps.
Your notes have 3 main topics:
- Pricing41%
- Onboarding33%
- Support26%
The gap: nobody connects onboarding to pricing.
Try asking: “What would onboarding look like if it explained the price?”
Check it yourself:open interview-notes in InfraNodus (opens in a new tab)
Just ask
Connect InfraNodus once. Then ask your AI for a graph of any file, chat or note, and it comes back with the topics, the gap, and a question worth exploring.
- 1Create a free account
- 2Copy your API key from the subscription page
- 3Add it to your AI below
- In Claude, open Settings > Connectors and add a custom connector.
- Paste the server address below and connect.
- Sign in with your InfraNodus API key from your subscription page.
https://mcp.infranodus.com- Open the Apps page, then Advanced settings, and turn on Developer mode.
- Click Create app, name it InfraNodus and paste the address below. Leave OAuth empty.
- Click Connect and enter your InfraNodus API key from your subscription page.
https://mcp.infranodus.com- Get your API key from your subscription page.
- Run this in your terminal, with your key in place of the placeholder.
claude mcp add infranodus -s user -- env INFRANODUS_API_KEY=your-api-key npx -y infranodus-mcp-server- In Obsidian, open Settings > Community plugins and turn them on.
- Install and enable the InfraNodus plugin.
- Paste your API key from your subscription page in Settings > InfraNodus > API key.
- Open your client's MCP settings (in Cursor: Settings > Tools and MCP) and edit the config file.
- Add this server with your API key from your subscription page.
{
"mcpServers": {
"infranodus": {
"command": "npx",
"args": ["-y", "infranodus-mcp-server"],
"env": { "INFRANODUS_API_KEY": "your-api-key" }
}
}
}Be the one who thinks deeply
One knowledge graph from first reading to published work. Start at the stage you are in.
For researchers and consultants
Read 40 interviews or a literature review as one picture
Start with a search query or topic of interest. You can also upload papers, transcripts or survey answers. Explore the themes and what connects them, and keep the best ideas in Project Notes, so your own framework grows with every project.
You bring
- search query
- papers.pdf
- interviews.txt
- survey.csv
If you have been asking
- How do I analyse 40 interview transcripts for themes?
- What is missing from my literature review?
For Obsidian users and note-takers
Find the connections your vault has not made yet
Upload your Markdown notes, see the topics and the gaps between them, develop an idea with AI, then bring it back into Obsidian with the plugin.
You bring
- notes.md
If you have been asking
- How do I find connections between my notes?
- How do I connect Claude to my Obsidian vault?
For AI and agent builders
Give your agents a memory with structure
Connect InfraNodus to Claude, Cursor or your own agents over MCP. Build a knowledge graph inside the conversation, then ask what your agents still miss about your codebase or docs.
You bring
- docs/*.md
- repo/
- claude-chat.txt
If you have been asking
- How do I give Claude memory that persists across sessions?
- Which MCP server can find gaps in my docs or codebase?
For SEO and content strategists
Write what search results and AI answers miss
Map what Google ranks and what AI answers say about a topic, compare it with what people search for, and turn the gap into the next content plan for you or a client.
You bring
- google-results.csv
- ai-answers.txt
If you have been asking
- What are my competitors covering that I am not?
- How do I get my site cited in AI answers?
Graphs of real texts
Not mock-ups. Each of these is a public graph you can open, made from a real book, a body of knowledge or a pile of customer reviews.
-
Ray Dalio's Principles
A whole management book as one map of ideas, used as a reasoning expert for an AI.
Open the live graph -
Traditional Chinese medicine
An ontology of a practice: how its concepts relate, and which ones rarely meet.
Open the live graph -
Reviews of a computer monitor
Hundreds of Amazon reviews read as one picture: what buyers praise, and what nobody mentions.
Open the live graph
Focused, yet adaptive. Rational, yet poetic.
Gregory Bateson coined a beautiful term: ecology of the mind. A mind that is ecological can see the whole, and it can zoom in on any idea. It embraces diversity, and it can obsess over one thing when it needs to. It discovers the obvious, and it ponders the gaps that have yet to be bridged.
I started working on InfraNodus in 2011 to help people think this way. Too often we get stuck in the same ideas and the same loops. The graph works like an antivirus for the mind: it shows what a text is focused on, what it leaves out, and where two ideas could meet.
It began as a research prototype, became a peer-reviewed method in 2019, and today researchers, writers, strategists and people building with AI use it every day.
Dmitry Paranyushkin
From the lab
We publish what we learn every week on the Nodus Labs channel: new workflows, honest comparisons and the occasional strong opinion.
-
Why your AI chat needs a knowledge graph
-
Better graph view than Obsidian
-
Your AI tools should be hard to use
Works where you think: ChatGPT and Claude via MCP, Obsidian, n8n, the API and a browser extension.
Think out loud with other people who use graphs
Join the Nodus Labs Discord. Ask how to read a graph, share what you found, and see what others are building with InfraNodus and AI.
Join the DiscordStart free, upgrade when it clicks
Sign up free, then pick a plan inside the app when you need more. Cancel or pause at any time. Student discounts and educational licenses are available.
Basic
€12
- Unlimited graphs and AI chats
- Files and PDFs up to 1 MB
- 40 AI credits per hour
- API and MCP (unlimited on annual)
Advanced
€32
- Everything in Basic
- Files and PDFs up to 5 MB
- 80 AI credits per hour
- Live updates and scheduled imports
Premium
€66
- Everything in Advanced
- PDFs up to 50 MB, files up to 10 MB
- 200 AI credits per hour
- Unlimited API and MCP, custom branding
Every plan starts free; upgrade in the app.
Sign up, then pick a planWhat is InfraNodus?
A visual AI text analysis tool that shows the patterns and gaps in any text. People use it for brainstorming, mapping a body of knowledge, market research and content strategy. Besides topic modelling and sentiment analysis, it represents your text as a knowledge graph: a map you can explore yourself, and context you can give to an AI for better answers. How it works
How does it compare to other tools?
Most tools are either too technical or too simple, and few use visualisation to explore data rather than just to summarise it. InfraNodus gives you insights usually found in advanced NLP tools in a visual interface, where the graph lets you zoom out for an overview and in for the relationships and gaps. It is built on an ecological thinking framework that helps you switch between focused and exploratory modes, and its API makes the same insights available to AI agents.
How do I use it?
Start with one of the import apps, for example Google search results for a topic. Then analyse your own texts: get an overview of the graph, explore the main topics and their statements, let the AI summarise them, and look for the gaps between topics. Finally, hide the concepts you have already analysed to see what lies underneath. The analysis workflow
How do I find the content gaps in my knowledge?
The graph shows the main topics in a discourse. Some of them are poorly connected: those are the content gaps, and they are where new ideas and insights come from. More on discovering content gaps
How is InfraNodus different from AI chatbots and analysis tools?
AI chatbots return the most likely answer, with reasoning you cannot see. InfraNodus uses AI-powered knowledge graphs instead.
Without InfraNodus
- Bullet points and long text
- Generic responses
- Reasoning hidden in a black box
- No graph, no deeper view of the context
With InfraNodus
- Reveals the patterns in your data
- Original insights from the structure
- An interactive graph shows what is happening
- GraphRAG finds the gaps in your knowledge, and the MCP server brings them to your LLM
How is InfraNodus different from traditional mind mapping?
Manual mind mapping is slow and biased towards what you already know. InfraNodus generates the graph from your text automatically.
Without InfraNodus
- Limited to hierarchical structures
- Manual creation is slow and prone to bias
- Does not reveal patterns and relations
- Few import sources
With InfraNodus
- A systematic approach to text analysis
- Knowledge graphs generated for you
- Network science reveals insights you would miss
- Import data from almost anywhere
How is InfraNodus different from standard text analysis tools?
Tag clouds and frequency tables hide the relationships. Text network analysis reveals them.
Without InfraNodus
- Tag clouds without context
- Sentiment only as positive or negative
- Results in tables that hide relationships
- Little room to explore
With InfraNodus
- Main topics, relations and content gaps
- Sentiment analysis together with topic modelling
- Relationships shown as a graph
- Metrics combined with exploration
Can I use it with ChatGPT, Claude or my own agents?
Yes. Connect the MCP server to Claude, ChatGPT, Cursor or n8n, or use the API. Your graphs then work as context and reasoning tools for the model.
Is my data private?
Yes. Your graphs are private by default, you can export them at any time, and they are hosted in the EU. Privacy policy
Your text, next
Create a free account, add one document, and see what it is missing.