How do I get and secure my LLM API keys in n8n?

You must obtain an API key directly from your chosen LLM provider (e.g., OpenAI, Google Cloud, Anthropic). Then, in n8n, navigate to 'Credentials,' create a new credential for the specific LLM service, and paste your API key there.

What is prompt engineering and why is it important for LLM integration?

Prompt engineering is the art and science of crafting effective text inputs (prompts) to guide an LLM to generate desired, relevant, and high-quality outputs. It involves clear instructions, context, examples, and format specifications.

How do I extract the generated text from an LLM node's output in n8n?

LLM outputs are typically structured as JSON objects. You can use n8n's 'Set' node with expressions (e.g., `{{ $json.choices[0].message.content }}`) or a 'Code' node to extract specific pieces of information, such as the generated text.

Are there costs associated with using LLMs in n8n, and how can I manage them?

LLM usage incurs costs based on token count (input and output) and model choice. Monitor your provider's pricing and consider using cheaper models for simpler tasks, optimizing prompts to reduce token usage, and implementing rate limits in n8n.

Back/AI Automation

Integrating Large Language Models (LLMs) in n8n: Your First AI Workflow

n8n

By Anuj SharmaJuly 22, 2026 • 3 MIN READ

The Brief

n8n integrates Large Language Models (LLMs) like OpenAI, Gemini, and Claude via dedicated nodes. Users configure API keys as credentials, then craft prompts within the LLM node to generate text, summarize, or classify data. Parsing the JSON output enables further workflow automation.

Action Checklist

Obtain an API key from your preferred LLM provider (e.g., OpenAI).
Create and securely save the LLM API key as a credential in your n8n instance.
Build a new n8n workflow incorporating an LLM node.
Configure the LLM node with your new credentials and a specific model.
Write a clear, concise prompt within the LLM node's message content.
Execute the workflow to observe the LLM's raw output.
Add a 'Set' node to parse and extract the generated text from the LLM's response.
Experiment with different prompts and models to understand their impact on output.

Key Takeaways

n8n provides native nodes for easy integration with leading LLMs like OpenAI, Gemini, and Claude.
Securely managing API keys as credentials is a critical first step for LLM integration.
Effective prompt engineering directly influences the quality and relevance of LLM outputs.
LLM outputs are typically JSON, requiring parsing to extract usable text for subsequent workflow steps.
Starting with basic text generation workflows is an excellent way to familiarize yourself with n8n's AI capabilities.
LLM integration in n8n unlocks powerful new possibilities for intelligent automation across various domains.

Having explored n8n's foundational components, AI concepts, and basic application integration, it's time to unlock the transformative power of artificial intelligence directly within your workflows. Large Language Models (LLMs) are central to modern AI automation. This chapter will guide you through connecting these powerful models to n8n. You will learn to craft effective prompts and interpret AI-generated responses. This is the first step towards building truly intelligent automations.

What Is It?

LLM integration in n8n refers to the process of connecting n8n workflows directly to Large Language Model APIs. This enables n8n to send text inputs (prompts) to an LLM, receive AI-generated text outputs, and then utilize these outputs for further automation. This integration leverages n8n's native LLM nodes, abstracting complex API calls into user-friendly graphical interfaces.

Why It Matters

Integrating LLMs into n8n workflows is crucial because it transforms traditional automation into intelligent automation. It allows systems to generate human-like text, summarize information, classify data, and even make context-aware decisions. This capability unlocks new use cases in content creation, customer support, data analysis, and beyond, significantly enhancing efficiency and reducing manual effort across various business functions.

When to Use It

Integrate LLMs in n8n when your workflow requires dynamic, context-aware text generation, summarization, translation, or classification. Use cases include drafting personalized marketing emails, generating social media captions, summarizing long articles, extracting specific entities from unstructured text, or building initial versions of AI chatbots. It is ideal for tasks needing human-like linguistic understanding and output.

Prerequisites

Chapter 1: Foundations of n8n and AI Automation(Understanding n8n's interface and core concepts)
Chapter 2: Core Concepts of Artificial Intelligence for Automation(Familiarity with LLMs and AI basics)
Chapter 3: n8n Interface and Basic Workflow Design(Ability to navigate n8n and create simple workflows)
Chapter 4: Integrating Applications and Services with n8n(Knowledge of setting up credentials and connecting to external services)

Step-by-Step Framework

Step 1: Access Your LLM Provider and Obtain an API Key. For OpenAI, visit platform.openai.com/api-keys. For Google Cloud (Gemini), follow their credential setup for API access. For Anthropic Claude, obtain a key from their console.

Step 2: Create New Credentials in n8n. In n8n, navigate to 'Credentials' in the left sidebar. Click 'New Credential' and search for your chosen LLM provider (e.g., 'OpenAI API'). Select it and paste your API key into the designated field. Give the credential a descriptive name (e.g., 'MyOpenAIKey') and save it securely.

Step 3: Start a New Workflow and Add an LLM Node. In the n8n canvas, create a new workflow. Add a 'Start' node. Search for your LLM provider's node (e.g., 'OpenAI') and drag it onto the canvas. Connect the 'Start' node to the 'OpenAI' node.

Step 4: Configure the LLM Node with Credentials and Model. Double-click the LLM node. In the 'Credentials' dropdown, select the credential you created in Step 2. Choose an appropriate 'Model' (e.g., 'gpt-3.5-turbo' for OpenAI, 'gemini-pro' for Google).

Step 5: Craft Your First Prompt. In the LLM node's parameters, locate the 'Messages' section. Add a new message. Set the 'Role' to 'User'. In the 'Content' field, enter your prompt. For example: 'Draft a concise, engaging social media post announcing a new online course on n8n AI automation. Highlight its benefits: flexibility, cost-effectiveness, and powerful integrations.'

Step 6: Execute the Workflow and Review Output. Click 'Execute Workflow' on the n8n canvas. After execution, inspect the output of the LLM node. The generated text will typically be within a 'data' array, often under 'choices[0].message.content' for OpenAI or similar paths for other providers.

Step 7: Parse the LLM Output. Add a 'Set' node after the LLM node. Use an expression like {{ $json.choices[0].message.content }} to extract the generated text. This isolates the desired content for use in subsequent nodes. Connect the 'Set' node to the LLM node and re-execute to see the parsed output.

Best Practices

Always store API keys as secure credentials in n8n, never hardcode them directly into nodes.

Start with simple, clear prompts; refine them iteratively to achieve desired outputs.

Specify the LLM model (e.g., GPT-3.5 Turbo vs. GPT-4) based on complexity and cost requirements.

Regularly test your LLM nodes with different prompts to understand their behavior and limitations.

Implement basic error handling for LLM nodes to manage API rate limits or invalid responses.

Parse LLM outputs using 'Set' or 'Code' nodes to ensure only relevant data is passed to downstream nodes.

Common Mistakes

Using incorrect or expired API keys, leading to authentication errors.

Crafting vague or ambiguous prompts, resulting in irrelevant or unhelpful LLM responses.

Forgetting to parse the LLM's JSON output, making the generated text inaccessible for subsequent steps.

Exceeding LLM rate limits or token limits without implementing retry mechanisms or chunking inputs.

Not selecting the appropriate LLM model for the task, leading to either overkill (expensive) or underperformance (poor quality).

Exposing API keys in logs or directly in node configurations, compromising security.

Recommended Tools & Resources

n8n: The core workflow automation platform.
OpenAI (GPT-3.5 Turbo, GPT-4): Leading LLM provider for diverse text generation tasks.
Google Gemini (Gemini Pro, Gemini Ultra): Google's advanced LLM for multi-modal and text-based AI.
Anthropic Claude (Claude 3 Haiku, Sonnet, Opus): Known for strong performance in complex reasoning and content generation.
Postman or similar API client: Useful for testing LLM APIs directly before implementing in n8n.

Frequently Asked Questions

n8n supports major LLM providers through dedicated nodes, including OpenAI (for GPT models), Google (for Gemini models), and Anthropic (for Claude models). Some community nodes also exist for other LLMs or self-hosted options.

Related Dispatches

Personal Brand

The Future of Personal Branding: Innovation & Ethical Considerations in the AI Age

Personal Brand

Advanced Personal Branding Frameworks: Scaling & Monetizing Your Influence

Next ChapterThe next chapter will build on this foundation by diving deeper into practical applications of LLM integration. We will explore specific workflows for AI-powered text generation and summarization, enabling you to automate content creation and distil information efficiently.

Action Checklist

Obtain an API key from your preferred LLM provider (e.g., OpenAI).

Create and securely save the LLM API key as a credential in your n8n instance.

Build a new n8n workflow incorporating an LLM node.

Configure the LLM node with your new credentials and a specific model.

Write a clear, concise prompt within the LLM node's message content.

Execute the workflow to observe the LLM's raw output.

Add a 'Set' node to parse and extract the generated text from the LLM's response.

Experiment with different prompts and models to understand their impact on output.

Key Takeaways

n8n provides native nodes for easy integration with leading LLMs like OpenAI, Gemini, and Claude.

Securely managing API keys as credentials is a critical first step for LLM integration.

Effective prompt engineering directly influences the quality and relevance of LLM outputs.

LLM outputs are typically JSON, requiring parsing to extract usable text for subsequent workflow steps.

Starting with basic text generation workflows is an excellent way to familiarize yourself with n8n's AI capabilities.

LLM integration in n8n unlocks powerful new possibilities for intelligent automation across various domains.