Cognee MCP. Find relationships, not just keywords.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Cognee. Build knowledge graphs from unstructured data. Cognee ingests raw text, extracts entities and relationships, and lets your AI agent search using graph-aware reasoning.
It's designed to go beyond standard search by mapping how facts connect, not just what they say. Use it to turn documents and reports into a structured, interconnected knowledge base your agent can query.
What your AI agents can do
Cognee add data
Ingests raw text or documents into the Cognee knowledge base. This must run before processing the data into a graph.
Cognee cognify
Processes ingested data, extracting entities and relationships to build a structured knowledge graph for querying.
Cognee get insights
Retrieves structured entity relationships and insights, showing how concepts connect across the knowledge base.
Sends raw text or documents to the knowledge base, making the data available for graph construction.
Processes ingested data to extract entities, map relationships, and build the structured graph required for advanced querying.
Retrieves structured entity relationships, showing how different concepts connect across the entire knowledge base.
Searches the knowledge graph using natural language, returning answers that trace connections using graph traversal and semantic search.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
019d7576cognee add data
Ingests raw text or documents into the Cognee knowledge base. This must run before processing the data into a graph.
019d7576cognee cognify
Processes ingested data, extracting entities and relationships to build a structured knowledge graph for querying.
019d7576cognee get insights
Retrieves structured entity relationships and insights, showing how concepts connect across the knowledge base.
019d7576cognee search
Queries the knowledge graph using natural language, returning context-aware answers via graph traversal and semantic search.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Cognee, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
You connect your AI agent to Cognee. This setup takes raw, unstructured text and turns it into a structured knowledge graph. Instead of just searching text, your agent can map how facts relate to one another. You use it to turn documents and reports into an interconnected knowledge base your agent can query.
Ingesting Raw Data
Use cognee_add_data to send raw text or documents straight into the knowledge base. This step makes the data available for graph construction.
Building the Knowledge Graph
Run cognee_cognify to process the ingested data. This tool extracts entities and maps relationships, building the structured graph you need for advanced querying.
Querying Structured Relationships
cognee_get_insights retrieves structured entity relationships, showing exactly how concepts connect across your entire knowledge base.
Performing Graph-Aware Search
Your agent uses cognee_search to query the knowledge graph with natural language. This process returns context-aware answers by tracing connections using graph traversal and semantic search.
How Cognee MCP Works
- 1 Use
cognee_add_datato feed the system raw text or documents you want to analyze. - 2 Run
cognee_cognifyto automatically extract entities and relationships, which converts the raw data into a graph structure. - 3 Call
cognee_searchorcognee_get_insightsto query the resulting knowledge graph using context-aware reasoning.
The bottom line is, you feed the data, build the map, and then ask questions that follow the map's connections.
Who Is Cognee MCP For?
This is for data architects and research analysts who are tired of flat, keyword-based search. If your business intelligence needs to understand why facts connect—not just that they exist—you need this. It's for anyone whose job involves synthesizing information from disparate documents.
Uses this to process dozens of academic papers or market reports, then asks the agent, 'What are the shared risks between these three companies?' to build a relationship map.
Builds a centralized, structured repository from internal wikis and operational logs, allowing agents to find complex connections across departments.
Implements the full pipeline: using cognee_add_data and cognee_cognify to normalize data sources before feeding it to downstream applications.
What Changes When You Connect
- Graph-Aware Search: Stop getting generic answers.
cognee_searchuses graph traversal, meaning it understands the connections between concepts, giving you answers that are rooted in the data's structure. - Structured Insights:
cognee_get_insightsshows you relationship maps. Instead of reading a list of facts, you get a visualization of how those facts relate to each other, which is crucial for complex research. - Hybrid Retrieval: This tool combines the best of two worlds: graph traversal (for structure) and vector search (for meaning). This hybrid approach gives superior recall compared to standard RAG systems.
- Full Data Lifecycle: The flow is baked in. You use
cognee_add_datato collect the data,cognee_cognifyto structure it, and thencognee_searchto use it. It handles the entire pipeline. - Temporal Awareness: Cognee tracks when facts were added. This means you can reason over time-based connections in your data, which plain vector databases miss.
Real-World Use Cases
Analyzing Cross-Company M&A Potential
A corporate development analyst needs to know if Company A's suppliers connect to Company B's key personnel. Instead of manually cross-referencing spreadsheets, they run cognee_add_data on all company reports. Then, they run cognee_cognify to map the connections. Finally, they use cognee_get_insights to get a structured view of shared suppliers and personnel, identifying M&A paths instantly.
Mapping Scientific Literature Connections
A biomedical researcher is sifting through hundreds of papers on gene interactions. They use cognee_add_data to ingest all PDFs. They run cognee_cognify to extract genes, proteins, and interaction types. They then use cognee_search to ask, 'What are the common regulatory mechanisms linking Protein X and Gene Y?' The agent returns a graph-aware answer citing specific connections.
Auditing Complex Internal Processes
The compliance officer needs to trace how a specific client record (Client ID 456) moved through the system over three years. They use cognee_add_data on all CRM logs. They run cognee_cognify to model the process steps and actors. They then use cognee_get_insights to visualize the full sequence and identify any gaps or non-standard handoffs.
Building a Product Knowledge Base
A technical writer gathers product specs, FAQs, and user manuals. They use cognee_add_data to upload everything. They run cognee_cognify to map product features, components, and use cases. Finally, they use cognee_search to build a chatbot that answers 'How does Component Z interact with Feature Q?' with accuracy.
The Tradeoffs
Simple document upload
You upload 50 PDFs and ask your agent, 'Tell me about the risks.' The agent only finds keywords 'risk' and 'liability' but can't tell you which company's risk relates to which product failure.
→
First, use cognee_add_data to ingest all 50 PDFs. Then, run cognee_cognify to build the graph. Finally, use cognee_search to ask, 'What is the relationship between the risk of product failure and the financial exposure of Company X?'
Relying only on keywords
Your agent returns a list of facts: 'Company A was founded in 2001. Company B was founded in 2010.' You can't tell if they are related or if they are just two random facts from the same document.
→
Run cognee_cognify first. This links the founding dates and companies. Then, use cognee_get_insights to explicitly map the relationship, like 'Company A was a competitor to Company B' using the graph.
Skipping the processing step
You try to query data immediately after uploading a file. The agent fails or gives vague results because the system hasn't built the underlying structure yet.
→
Always follow the sequence: cognee_add_data → cognee_cognify → cognee_search. Don't skip the graph building step.
When It Fits, When It Doesn't
Use this if your goal is to build a knowledge base where understanding relationships is the primary goal. If you need to know how Concept A influences Concept B, or if you need to track a complex, multi-step process, Cognee is for you. Don't use it if you just need to search simple facts (e.g., 'What is the capital of France?'). For simple lookups, a standard vector search tool works fine. If your data is messy and comes from diverse sources (PDFs, logs, reports), you need the full pipeline: cognee_add_data for ingestion, cognee_cognify for structure, and cognee_search for reasoning. If your data is already perfectly structured in a database, you don't need Cognee; you need a direct SQL-style interface.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Cognee. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 4 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Manually stitching together insights from dozens of documents is a nightmare.
Today, if you need to know the relationship between a supplier's location and a product's failure rate, you open a dozen different documents. You copy-paste the key facts into a spreadsheet, then you manually try to connect the dots. It takes hours, and you always miss some critical contextual link.
With Cognee, you dump all those documents into the server. You run the `cognee_cognify` tool. The system automatically extracts every person, company, and product, and maps the relationship between them. You don't copy, you just ask.
Use `cognee_search` to get graph-aware answers.
Before, you'd search for 'product failure' and get a list of documents that mention the words. You'd have to read through them all to figure out who was responsible, or what the contributing factor was.
Now, using `cognee_search`, you ask the question, and the agent returns an answer that shows the connections. It doesn't just say 'yes'; it says 'yes, and here is the path it took, connecting A to B via C'.
Common Questions About Cognee MCP
How do I use the `cognee_add_data` tool? +
You use cognee_add_data to feed the system the raw text or documents. This is the initial step that makes the data available. You can't run any other tool until the data is ingested here.
What is the difference between `cognee_search` and `cognee_get_insights`? +
cognee_search gives you a direct, answer-based response to a question. cognee_get_insights gives you a structured view of relationships, showing how concepts connect across the entire knowledge graph.
Do I need to run `cognee_cognify` every time I add new data? +
Yes, you must run cognee_cognify after adding data. This process extracts the entities and relationships, building the actual graph structure. Without it, the data is just raw text, not knowledge.
Can I query my data using `cognee_search` without first running `cognee_cognify`? +
No. The graph-aware search relies entirely on the structured relationships created by cognee_cognify. If you skip that step, the search tool won't find the connections you need.
How does `cognee_cognify` handle data I've already processed? +
It intelligently updates the graph structure. You don't have to rebuild everything; the tool detects existing entities and relationships, merging new data points without overwriting old knowledge.
What are the prerequisites for using `cognee_search`? +
You must first run cognee_cognify on your data. The search tool relies on the structured graph built by cognee_cognify to provide context-aware answers.
Can I use `cognee_get_insights` on multiple datasets at once? +
Yes, you can structure the query to include multiple datasets. cognee_get_insights then compiles a unified view, showing how concepts connect across your entire knowledge base.
What kind of data is best suited for `cognee_add_data`? +
Raw, unstructured text or document files work best. This is the intended input for cognee_add_data, which prepares the data for entity extraction.
How is Cognee different from standard RAG? +
Standard RAG splits documents into chunks and finds similar text using vector search — but it loses the relationships between facts. Cognee builds a knowledge graph that preserves entity relationships, temporal connections, and hierarchical structures. When you search, Cognee uses graph traversal combined with vector similarity and LLM reasoning, resulting in more accurate, context-aware answers that understand HOW facts relate to each other.
What search types are available? +
Cognee supports four retrieval strategies: GRAPH_COMPLETION (default — combines vector search + graph traversal + LLM reasoning for context-aware answers), SUMMARIES (fast hierarchical overview search), INSIGHTS (structured entity relationships), and CHUNKS (pure vector similarity for raw text passages). Each strategy optimizes for different use cases.
Is Cognee open-source? +
Yes! Cognee is fully open-source under the Apache 2.0 license. You can self-host the entire platform including the knowledge graph engine, vector database, and API server. A managed cloud version with API keys is also available for teams that prefer not to manage infrastructure.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Scale AI
Automate high-quality data labeling and RLHF workflows via Scale AI — create projects, manage batches, and submit annotation tasks directly from your AI agent.
Cohere (AI Platform)
Power enterprise AI via Cohere — generate text, perform chat completions, reorder documents, and manage embeddings directly from any AI agent.
Leonardo.ai (Generative AI & Models)
Generate high-fidelity images via Leonardo.ai — orchestrate generations, audit AI models, and manage visual assets.
You might also like
Mailchimp
Synchronize marketing contacts — list audiences, control subscriber limits, and edit campaign vectors autonomously.
Chiirp
Engage leads and customers through automated text, email, and voicemail campaigns that nurture relationships at scale.
Moova
Automate smart logistics via Moova — create shipments, track deliveries, get shipping quotes, and orchestrate carriers from any AI agent.