> For the complete documentation index, see [llms.txt](https://graphai-1.gitbook.io/graphai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://graphai-1.gitbook.io/graphai/how-graphai-works/core-components-of-graphai.md).

# Core Components of GraphAI

The GraphAI workflow consists of several key stages:

1. Data Ingestion: Continuous indexing of blockchain data
2. Sub-Index Creation: User-defined creation of specialized data subsets
3. Vector Database Storage: Efficient storage of sub-indexes for AI operations
4. Knowledge Graph Integration: Mapping relationships and context within the data
5. MCP Access: Enabling developers to query and utilize the processed data
6. AI/LLM Application Development: Creation of intelligent dApps leveraging Graphai's capabilities

### Blockchain Data Indexing

GraphAI implements a comprehensive indexing system that processes both historical and real-time blockchain data. This system allows for efficient querying and retrieval of blockchain information, forming the foundation for all subsequent data processing and AI applications.

### Customisable Sub-Indexes

Developers and users can create tailored sub-indexes based on specific criteria or data points within the blockchain. These sub-indexes enable focused and efficient data retrieval for specialized applications, significantly reducing query times and computational overhead.

### Vector Database Integration

Sub-indexes are stored in high-performance vector databases, optimizing them for AI and machine learning operations. This storage method allows for rapid similarity searches and efficient processing of high-dimensional data, crucial for many AI applications.

### Knowledge Graph Construction

GraphAI incorporates a dynamic knowledge graph that maps relationships and contexts within the blockchain data. This graph enhances the semantic understanding of the data, enabling more intelligent and context-aware AI applications.

GraphAI leverages GraphRAG technology to create a rich, interconnected representation of blockchain data. This knowledge graph goes beyond simple indexing by establishing semantic relationships between different data points, transactions, and blockchain states.

### Contextualising Blockchain Data

GraphRAG allows GraphAI to provide context to various types of blockchain data:

* Transactional Data: Linking related transactions, identifying patterns, and establishing transaction chains.
* Stateful Data: Mapping the evolution of smart contract states and account balances over time.
* Metadata: Connecting off-chain metadata with on-chain actions for a more comprehensive view.
* Temporal Relationships: Establishing timelines and causal relationships between blockchain events.

### Enhancing AI Model Undertanding

By utilizing GraphRAG, GraphAI enables AI models, particularly LLMs, to gain a deeper understanding of blockchain data context. This enhanced comprehension allows for more accurate and nuanced interactions with blockchain information.