LLM Prompt

Generate AI-powered text responses using Large Language Models

Node Type

Action

Icon

Brain

Overview

The LLM Prompt node is an action node that sends prompts to Large Language Models and returns their responses. This powerful AI integration enables AI-powered text generation, analysis, and processing within workflows, perfect for creating dynamic content, analyzing text, or generating creative responses.

Key Features

• AI-Powered Generation: Uses Large Language Models for intelligent text responses
• Flexible Prompting: Send any text prompt to the LLM for processing
• Temperature Control: Adjust creativity vs. determinism in responses
• Vision Analysis: Include image URLs for multimodal AI processing
• Cost Tracking: Monitor API usage costs for budget management
• HTML Output: Generate formatted HTML content for emails and web applications

Prerequisites

AI Service Access

Must have access to Large Language Model services

LLM service access through NodeServiceRegistry

Valid service credentials and API access

Sufficient API credits for LLM operations

Content Requirements

Prompt Design: Ability to write clear, effective prompts for the LLM

Output Formatting: Understanding of desired response formats (HTML, text, etc.)

Content Strategy: Clear understanding of what you want the AI to generate

Technical Requirements

LLM Integration: Access to LLM service through NodeServiceRegistry

Network Access: Internet connectivity for AI service communication

Error Handling: Proper exception handling for API failures

Node Configuration

Required Fields

Prompt

Type:text

Required:Yes

Value Type:string

The text prompt to send to the Large Language Model. This should clearly describe what you want the AI to generate, analyze, or process. For HTML output, include specific formatting instructions.

Optional Fields

Temperature

Type:number

Required:No

Value Type:number

Controls the randomness of the LLM's response. Higher values (closer to 1) make the AI more creative and unpredictable, while lower values (closer to 0) make responses more deterministic and focused.

Image URLs

Type:text

Required:No

Value Type:array_string

Optional array of image URLs to include with the prompt for vision analysis. This enables multimodal AI processing where the LLM can analyze both text and visual content together.

Best Practices

Do's

• Be specific and clear in your prompts for better results
• Use appropriate temperature settings for your use case
• Include formatting instructions when you need specific output formats
• Test different prompt variations to find what works best
• Use template variables for dynamic content generation
• Monitor API costs and usage patterns
• Provide context and examples in your prompts when helpful

Don's

• Don't use overly vague or ambiguous prompts
• Avoid extremely high temperatures for critical business content
• Don't forget to specify output format requirements
• Avoid prompts that could generate inappropriate content
• Don't ignore cost tracking and API usage limits
• Avoid overly complex prompts that may confuse the AI
• Don't assume the AI will understand industry jargon without context

💡

Pro Tip: When generating HTML content, always include specific formatting instructions in your prompt. For emails, specify that you want "nicely formatted visually appealing HTML" and instruct the LLM to output only raw HTML without markdown formatting or code blocks.

Troubleshooting

Common Issues

Service Connection Failures

Symptoms: Node fails with service connection or registry errors

Solution: Verify that the LLM service is properly registered in the NodeServiceRegistry and that all required credentials and configurations are set up correctly.

Poor Response Quality

Symptoms: LLM responses are irrelevant or low quality

Solution: Improve your prompt by being more specific, providing context, and using appropriate temperature settings. Test different prompt variations to find what works best.

HTML Formatting Issues

Symptoms: Generated HTML includes markdown or code blocks

Solution: Include explicit instructions in your prompt: "Output only the HTML, not markdown. Do NOT output ```html or ```markdown. Just output the raw HTML."

High API Costs

Symptoms: Unexpectedly high costs from LLM API usage

Solution: Monitor the cost output from the node, optimize prompts to be more concise, and use appropriate temperature settings. Consider implementing cost controls in your workflows.