Build an AI Agent That Picks the Best AI Model (n8n Tutorial)

Key Concepts

• AI Agent
• Large Language Models (LLMs)
• Langchain
• Nodes (AI Agent Node, Langchain Code Node)
• Cost-effective AI workflows
• Model Switching
• Local Hosting vs. Cloud Hosting
• VPS (Virtual Private Server)

Main Topics and Key Points

• Introduction to AI Agents with Multiple LLMs: The video demonstrates how to create an AI agent within N8N that can dynamically switch between multiple large language models (LLMs) or chat models. This is beneficial for optimizing cost and speed in various workflows.
• Customer Complaint Workflow Example: The video references a customer complaint workflow (originally by Mario) as an example. This workflow takes a customer complaint, generates a response, validates it, and regenerates it with a different (potentially larger) model if the initial response is unsatisfactory. The link to the workflow is provided in the video description.
• Limitations of the Native N8N AI Agent Node: The standard AI Agent node in N8N only allows the use of one chat model at a time. To enable dynamic switching between models, a Langchain code node is required.
• Langchain as the Foundation: Langchain is described as a popular framework and the building block for AI agents. It provides the ability to interact with multiple tools and LLMs. The video references a beginner's course within the creator's community that explains Langchain's connection to N8N, chain nodes, and the AI Agent node.
• Using the Langchain Code Node:
  • The video details how to add a Langchain code node to an N8N workflow by navigating to "Add Node" -> "AI" -> "Other AI Nodes" -> "Miscellaneous" -> "Langchain Code Node."
  • The Langchain code node is initially a plain node. To enable LLM and tool integration, inputs must be added.
  • Adding an "input" of type "Language Model" to the Langchain code node allows the user to add multiple LLMs (e.g., OpenAI, Grok, Anthropic).
  • Adding an "input" of type "Tools" enables the use of various tools within the Langchain code node.
• Local Hosting Requirement: The Langchain code node with multiple LLM support is only available when N8N is hosted locally or on a VPS (Virtual Private Server), not on the cloud version of N8N.
• Hostinger VPS Setup: The video recommends Hostinger as a cost-effective VPS option for hosting N8N. The KVM2 plan is suggested.
  • The video provides a step-by-step guide on setting up N8N on Hostinger:
    1. Choose the KVM2 plan.
    2. Select a billing period (12 or 24 months recommended for better discounts).
    3. Apply the coupon code "AIWORKSHOP" for an additional 10% discount.
    4. Search for "N8N" in the OS search.
    5. Click "Confirm" and then "Continue."
    6. Register or log in to Hostinger.
    7. Add billing and payment information.
    8. Click "Install."
    9. After installation, click "Manage App" to access the N8N dashboard.
    10. Click "Manage App" again to open the N8N workflow.
• Custom Code Integration: The video demonstrates how to add custom code to the Langchain code node. This allows for specific logic, such as switching between models based on an index within a loop. The example code used in the customer complaint workflow is shown.

Important Examples, Case Studies, or Real-World Applications Discussed

• Customer Complaint Workflow: This workflow serves as a practical example of how to use multiple LLMs to generate and validate responses, switching to a more powerful model if the initial response is unsatisfactory.

Step-by-Step Processes, Methodologies, or Frameworks Explained

• Setting up N8N on Hostinger VPS: The video provides a detailed, step-by-step guide on how to set up N8N on a Hostinger VPS, including selecting a plan, applying a coupon code, and installing N8N.
• Creating a Langchain Code Node with Multiple LLMs: The video outlines the process of adding a Langchain code node to an N8N workflow and configuring it to use multiple LLMs and tools.

Key Arguments or Perspectives Presented, with Their Supporting Evidence

• Cost-Effectiveness of Model Switching: The video argues that using an AI agent capable of switching between LLMs can lead to more cost-effective AI workflows by starting with smaller, cheaper models and only using larger models when necessary.
• Importance of Local Hosting for Advanced Features: The video emphasizes that certain advanced features, such as the Langchain code node with multiple LLM support, are only available when N8N is hosted locally or on a VPS.

Notable Quotes or Significant Statements with Proper Attribution

• "Every AI agent inside your N8N is an instance of a Langchain."
• "Langchain is one of the most popular frameworks out there in the market when it comes to AI agent. It's basically the building blocks of AI agent because it gives you the ability to add like you know interact with multiple tools and everything else."
• "This only uh works on a local host or a VPS."

Technical Terms, Concepts, or Specialized Vocabulary with Brief Explanations

• AI Agent: An autonomous entity that can perceive its environment, make decisions, and take actions to achieve specific goals.
• Large Language Model (LLM): A deep learning model trained on a massive dataset of text data, capable of generating human-like text, translating languages, and answering questions. Examples include GPT-4, Claude, and Grok.
• Langchain: A framework for building applications powered by language models. It provides tools and abstractions for connecting LLMs to various data sources and tools.
• Node: A building block in an N8N workflow that performs a specific task, such as connecting to an API, transforming data, or executing code.
• VPS (Virtual Private Server): A virtual machine that provides dedicated resources and operating system, allowing users to host applications and websites with greater control and flexibility compared to shared hosting.

Logical Connections Between Different Sections and Ideas

The video starts by introducing the concept of AI agents with multiple LLMs and then explains why this is useful. It then highlights the limitations of the native N8N AI Agent node and introduces Langchain as the solution. The video then provides a step-by-step guide on how to create a Langchain code node with multiple LLMs and tools. Finally, it emphasizes the importance of local hosting and provides a detailed guide on setting up N8N on a Hostinger VPS. The customer complaint workflow serves as a practical example throughout the video.

Data, Research Findings, or Statistics Mentioned

• Hostinger VPS cost: approximately $6 per month (before discounts).
• Additional 10% discount available with the coupon code "AIWORKSHOP" on Hostinger.

Brief Synthesis/Conclusion of the Main Takeaways

The video provides a practical guide on creating more efficient and cost-effective AI workflows in N8N by leveraging the Langchain code node to dynamically switch between multiple LLMs. It emphasizes the importance of local hosting or using a VPS like Hostinger to access this functionality and provides a detailed walkthrough of the setup process. The customer complaint workflow serves as a compelling example of how this approach can be applied in real-world scenarios. The key takeaway is that by using Langchain and strategically switching between LLMs, users can optimize their AI workflows for both performance and cost.