n8n Workflow: Build Comprehensive Entity Profiles with GPT-4, Wikipedia & Vector DB

This n8n workflow demonstrates how to leverage AI (GPT-4), public knowledge (Wikipedia), and a vector database (Qdrant) to build comprehensive profiles for entities. It's designed to enrich incoming data by fetching relevant information and structuring it for further use.

What it does

This workflow performs the following key steps:

1. Triggers Manually or via External Workflow: The workflow can be initiated manually with a click or by another n8n workflow, allowing for flexible integration into larger automation pipelines.
2. Initial Data Preparation: An "Edit Fields (Set)" node is used to prepare the incoming data, likely setting or transforming initial parameters for the entity profile generation.
3. Conditional Logic: An "If" node introduces conditional branching, enabling the workflow to execute different paths based on specific criteria in the data. This allows for dynamic processing.
4. AI Agent for Intelligent Processing: An "AI Agent" node (LangChain Agent) is at the core of the intelligent processing, likely orchestrating the use of various tools and language models to gather and synthesize information.
5. Leverages OpenAI Chat Model: The "OpenAI Chat Model" is utilized by the AI Agent for sophisticated natural language understanding, generation, and reasoning, likely powered by GPT-4.
6. Integrates Wikipedia for Knowledge Retrieval: A "Wikipedia" tool is employed by the AI Agent to fetch factual information about the entities, ensuring the profiles are grounded in reliable public knowledge.
7. Calls External n8n Workflows (Tool): The "Call n8n Workflow Tool" allows the AI Agent to trigger other n8n workflows as sub-tasks, enabling modularity and reusability of logic.
8. Loads Document Data: A "Default Data Loader" prepares documents or text for processing, likely for chunking or embedding.
9. Splits Text for Processing: A "Character Text Splitter" breaks down large texts into smaller, manageable chunks, which is crucial for efficient processing by language models and vector databases.
10. Generates Embeddings with Ollama: The "Embeddings Ollama" node creates vector embeddings from the text chunks. These numerical representations capture the semantic meaning of the text.
11. Stores and Retrieves Data from Qdrant Vector Store: A "Qdrant Vector Store" is used to store the generated embeddings. This allows for efficient semantic search and retrieval of relevant information based on similarity.
12. Structures AI Output: A "Structured Output Parser" processes the output from the AI agent, ensuring it conforms to a predefined structure (e.g., JSON), making it easy to consume by subsequent nodes or systems.
13. Merges Data: A "Merge" node combines data from different branches of the workflow, bringing together all the gathered and processed information into a unified output.
14. Provides Explanatory Notes: "Sticky Note" nodes are used throughout the workflow for documentation, providing context and explanations for specific sections or nodes.

Prerequisites/Requirements

• n8n Instance: A running n8n instance.
• OpenAI API Key: For the "OpenAI Chat Model" (likely GPT-4).
• Ollama Instance: For the "Embeddings Ollama" node (requires a running Ollama server with the desired embedding model).
• Qdrant Instance: A running Qdrant vector database instance.
• LangChain Integration: The n8n LangChain nodes package (@n8n/n8n-nodes-langchain) must be installed on your n8n instance.

Setup/Usage

1. Import the Workflow: Download the JSON provided and import it into your n8n instance.
2. Configure Credentials:
   • Set up your OpenAI API Key credentials for the "OpenAI Chat Model" node.
   • Configure the connection details for your Ollama instance in the "Embeddings Ollama" node.
   • Provide the necessary connection details (e.g., host, API key) for your Qdrant instance in the "Qdrant Vector Store" node.
3. Review Node Configurations:
   • Examine the "Edit Fields (Set)" node to understand how initial data is processed.
   • Adjust the conditions in the "If" node as per your specific branching logic.
   • Inspect the "AI Agent" node to understand its prompt, tools, and memory configuration.
   • Verify the settings for the "Wikipedia" tool, "Call n8n Workflow Tool" (if used to call other workflows), "Character Text Splitter", and "Structured Output Parser" to match your requirements.
4. Activate the Workflow: Once configured, activate the workflow.
5. Execute the Workflow:
   • You can execute it manually by clicking "Execute Workflow" in the n8n UI.
   • Alternatively, trigger it from another n8n workflow using the "Execute Workflow Trigger" node, passing the necessary input data.

This workflow provides a robust framework for building intelligent entity profiling systems, combining the power of large language models with specialized data sources and vector search capabilities.