Understanding Edges
Processing Edges
LLM Edge

LLM Edge

Use this edge to interact with large language models (LLMs) via API or locally hosted instances.

LLM_Edge

Purpose:

  • Generate text outputs from structured or unstructured input.
  • Answer questions, summarize, or rewrite content.

Inputs:

  • A text block or structured block with relevant information.

Outputs:

  • A text block with the LLM's response.

Configuration:

  • Messages: Prompt that include system and user roles. You can refer to input blocks using double curly braces (e.g. {{block_label}}). Click to copy the input block’s label and paste in the prompt.
  • Output type: Default set to text, can choose to enable structured output to let the LLM output contents in a JSON object.
  • Model: Select a good model to use.
  • Settings: Token limit and Base URL.

Available Models:

  • OpenAi: o1-mini , o1 , o1-pro , o3-mini , o3-mini-high , gpt-4-turbo , gpt-4.5-preview , gpt-4o-mini , gpt-4o
  • Anthropic: claude-3.5-haiku , claude-3.5-sonnet , claude-3.7-sonnet
  • DeepSeek: deepseek-v3 , deepseek-r1
  • Localhost: Any model that has depolyed in your local server and has connected to PuppyAgent.
OverviewModify Edge