LLM Node

​

Node Details

• Name: seqLLMNode

• Type: LLMNode

• Category: Sequential Agents

• Version: 3.0

​

Parameters

1. Name (string)

   • Label: Name

   • Description: A unique identifier for the LLM Node

   • Required: Yes

2. System Prompt (string)

   • Label: System Prompt

   • Description: Initial prompt to set the context for the LLM

   • Optional: Yes

   • Additional Parameter: Yes

3. Human Prompt (string)

   • Label: Human Prompt

   • Description: Prompt added at the end of the messages as a human message

   • Optional: Yes

   • Additional Parameter: Yes

4. Messages History (code)

   • Label: Messages History

   • Description: List of messages between System Prompt and Human Prompt, useful for few-shot examples

   • Optional: Yes

   • Additional Parameter: Yes

5. Sequential Node (list)

   • Label: Start | Agent | Condition | LLM | Tool Node

   • Description: Predecessor nodes in the sequential workflow

   • Required: Yes

6. Chat Model (BaseChatModel)

   • Label: Chat Model

   • Description: Overwrite model to be used for this node

   • Optional: Yes

7. Format Prompt Values (json)

   • Label: Format Prompt Values

   • Description: Assign values to prompt variables, can use $flow.state.variable-name

   • Optional: Yes

   • Additional Parameter: Yes

8. JSON Structured Output (datagrid)

   • Label: JSON Structured Output

   • Description: Instruct the LLM to give output in a JSON structured schema

   • Optional: Yes

   • Additional Parameter: Yes

9. Update State (tabs)

   • Label: Update State

   • Description: Mechanism to update the state after LLM execution

   • Optional: Yes

   • Additional Parameter: Yes

   • Tabs:

     • Update State (Table): UI-based state update

     • Update State (Code): Code-based state update

​

Input

The LLM Node receives input from its predecessor nodes in the sequential workflow. This can include:

• Previous agent outputs

• Tool results

• Condition evaluations

• Initial input from the Start node

​

Output

The LLM Node produces an output that typically includes:

• Content: The main response from the L