Overview

The Call LLM node is the core component for interacting with Large Language Models in HyperFlow. As a super-node, it provides a unified interface to dozens of different LLM service implementations, processing user inputs through various LLMs for text, image, video, or audio generation, supporting template-based prompts, knowledge extraction, and configurable generation parameters.

This node serves as the primary vehicle for implementing LLM-based agents in agentic workflows, including plan generation and execution agents, data-flow analyzers for guardrail components, content generation engines, code generators, and other autonomous decision-making components. It is the foundational building block for both conversational AI applications and sophisticated multi-agent systems.

Node Information

Purpose

Use the Call LLM node to:

• Generate text responses in conversational workflows
• Create images, videos, or audio content using AI
• Process user queries with RAG (Retrieval Augmented Generation)
• Build multi-turn conversations with context awareness
• Accept knowledge segments for in-context learning in RAG applications
• Automatically display cited knowledge metadata in connected chatbots
• Transcribe audio content to text

Agentic Applications

• Plan Generation: Create autonomous agents that generate execution plans
• Plan Execution: Implement agents that follow and execute structured plans
• Data Analysis: Deploy analyzers that examine data flows and provide insights
• Guardrails: Add safety components that validate and filter content
• Code Generation: Build agents that generate, review, or modify code