AI Text

Sends a prompt to an AI language model and returns the text response. Use this node for tasks like summarisation, classification, data extraction, translation, or any prompt-in / text-out operation.

Properties

Property	Required	Description
Model	No	The language model to use. Defaults to the model configured in your deployment settings
System Prompt	No	Instructions that shape how the model responds (its "role" or persona). Supports expressions
User Message	Yes	The prompt to send to the model. Supports expressions

Model options

Option	Model ID
Default (configured)	Uses the model set in OpenRouter.Model in your deployment config
Claude 3.5 Sonnet	anthropic/claude-3.5-sonnet
Claude 3 Haiku	anthropic/claude-3-haiku
GPT-4o	openai/gpt-4o
GPT-4o Mini	openai/gpt-4o-mini
Gemini 2.0 Flash	google/gemini-2.0-flash-001
Llama 3.3 70B	meta-llama/llama-3.3-70b-instruct
DeepSeek V3	deepseek/deepseek-v3

Models are served via OpenRouter. Your deployment's API key must have credits for the selected model.

Output

Field	Type	Description
text	String	The model's response
model	String	The model ID that was actually used

Access these in downstream nodes via expressions:

{{ $nodes['Summarise'].text }}
{{ $nodes['Summarise'].model }}

Replace Summarise with the actual name you gave the node.

Examples

Summarise an HTTP response

• System Prompt: You are a concise technical writer. Respond with a single paragraph.
• User Message: Summarise the following API response: {{ $nodes['Fetch Article'].responseBody }}

Classify a support ticket

• User Message: Classify the following support message into one of: billing, technical, general. Respond with only the category name.\n\n{{ $nodes['Webhook'].body.message }}

Extract structured data

• System Prompt: You extract data from text and respond only with valid JSON. No markdown.
• User Message: Extract the company name, date, and total amount from this invoice text:\n\n{{ $nodes['Parse PDF'].text }}

Use a Transform node downstream to map the extracted fields into your data model.