Basic LLM Chain

Send a prompt to a chat model and get its answer. The simplest way to use AI in a workflow.

Try it

workflow · explain-eli5.awf 4 steps

Explain Like I'm 5: a workflow with 4 steps, starting with Right-click a selection and ending with Show the explanation.

1. 1 Right-click a selection(Context Menu — open its doc page) Context Menu
2. 2 Get selected text(In-page action — open its doc page) In-page action
3. 3 Explain simply(Basic LLM Chain — open its doc page) Basic LLM Chain + Local AI model
4. 4 Show the explanation(Display Markdown — open its doc page) Display Markdown

Open in AWFlow Copy .awf

What it does

Basic LLM Chain sends Prompt to the chat model connected to its Model slot and returns the reply as { response }. There are no tools, no memory and no document lookup: one prompt in, one answer out.

You must connect a chat model, for example OpenAI, Anthropic, Ollama or an on-device Web LLM model. Optionally connect an output parser to get structured data back instead of plain text.

When to use it

• Rewrite, translate, classify or summarise text from earlier nodes.
• Draft an email, a reply or a title from page data.
• Turn free text into fields, with a Structured Output Parser attached.

When the model must call tools or take actions, use Tools Agent. To answer from your own documents, use RAG Agent or Q&A Agent.

Inputs and settings

Setting	Type	What it does
Prompt REQ(required)	Text
Options	Group	The options for the node.
Options › System message	Long text	The system message to use.

Use expressions in Prompt to include data, for example Write a one-line summary of: {{ $input.text }}. Add a System message under Options to set tone or rules (“Answer in French”, “Reply in under 50 words”).

Outputs

• Output

Each output item has these fields:

Field	Type	What it holds
response	Any	The model reply; the parsed result when an output parser is connected

Shape of an output item

{
  "response": null
}

response is the model’s reply: text, or the structured data your output parser produced.

Dependencies and credentials

Slot	Accepts	Connections
Model REQ(required)	chat model	One node.
Output Parser	output parser	One node.

The node has no credential of its own. The connected model holds the API key or connection.

Example workflow

1. Get Selected Text reads what you highlighted, with Output field name text.
2. Basic LLM Chain with Prompt Explain this in plain words: {{ $input.text }}, and an OpenAI model connected to Model.
3. Display Markdown shows {{ $input.response }}.

Common issues

• “Dependency model is required”. Connect a chat model to the Model slot.
• “Model not found”. The connected model node is not set up: open it and choose a model and its credential.
• The reply is not in the shape you need. Attach a Structured Output Parser, or ask for the format in the System message.

Recipes with this node

Explain any highlighted text in plain words beginner · 4 min On-device Rewrite rough text in a professional tone beginner · 4 min On-device

Related nodes

• Tools Agent: a model that can call tools.
• Information Extractor: pull named fields out of text.
• Summarization: summarise long text.
• Chrome AI: a prompt to Chrome’s built-in model, no model node needed.