Chat Model Prompter

{
    "title": "number_adder",
    "type": "object",
    "description": "Adds two numbers.",
    "properties": {
        "a": {
            "title": "A",
            "type": "integer",
            "description": "First value to add"
        },
        "b": {
            "title": "B",
            "type": "integer",
            "description": "Second value to add"
        }
    },
    "required": ["a", "b"]
}

Options

System message

Optional instructional message provided to the model at the start of the conversation, usually used to define various guidelines and rules for the model to adhere to.

Note: Certain models don't support system messages (e.g. OpenAI's o1-mini). For such models, the system message should be left empty.

Example: You are an expert in geospacial analytics, and you only reply using JSON.

New message

Optional next message to prompt the chat model with. If provided, a corresponding row will be appended to the conversation table.

Output format

Choose between different output formats.

Available options:

• Text: Text output message generated by the model.
• JSON: When JSON is selected, the model is constrained to only generate strings that parse into valid JSON object. Make sure you include the string "JSON" in your prompt or system message to instruct the model to output valid JSON when this mode is selected. For example: "Tell me a joke. Please only reply in valid JSON." Please refer to the OpenAI guide to see which models currently support JSON outputs.

Extend existing conversation

If enabled, messages produced by this node will be appended to the provided conversation table.

Otherwise, the output table will only contain the new message and the model's reply.

Note: If an existing conversation is provided, it will still be used as context if this setting is disabled.

Conversation

Role column

Select the column of the conversation table that specifies the role assigned to each message. The column can be empty if starting the conversation from scratch.

Example roles: 'human', 'ai'.

Message column

Select the column of the conversation table that specifies the messages. The column can be empty if starting the conversation from scratch.

Ignore new message during tool calling

If enabled, the 'New message' specified in the configuration dialog will not be appended to the conversation table while the last row is a tool call message.

In most cases, the new message would be what caused the model to invoke tool calling in the first place, and will have already been appended to the conversation table.

Tool name column (‘Conversation History’ table)

Select the column of the conversation table specifying tool names as strings.

This column gets populated with the name of tool the model decides to call.

Tool call ID column (‘Conversation History’ table)

Select the column of the conversation table specifying tool call IDs as strings.

This column gets populated with IDs of tool calls, so that they can be referenced by the model.

Tool call arguments column (‘Conversation History’ table)

Select the column of the conversation table specifying tool call arguments as JSON objects.

This column gets populated with the expected input arguments for the tool the model decides to call.

Tool definition column (‘Tool Definitions’ table)

Select the column of the tool definitions table containing definitions of the tools that should be available to the chat model.

Tool definitions take the form of JSON Schema-based objects, specifying the tool's name, description, parameters, and required fields.

Views

This node has no views

Workflows

• No workflows found

Links

• No links available

Developers