Ask Assistant

Action ID: openai_ask_assistant

Description

Ask a GPT assistant anything you want! This node allows you to interact with OpenAI assistants, which are specialized AI agents configured with specific instructions, tools, and knowledge. You can ask questions and maintain conversation memory across runs using memory keys.

Provider

OpenAI

Connection

Name	Description	Required	Category
OpenAI Connection	The OpenAI connection to use for the assistant.	✓	openai

Input Parameters

Name	Type	Required	Default	Description
assistant	dropdown	✓	-	The assistant which will generate the completion.
prompt	string	✓	-	The text prompt to ask the assistant.
memory_key	string	-	-	A memory key that will keep the chat history shared across runs. Leave empty to leave your assistant without memory of previous messages.

View JSON Schema

{
  "description": "Ask Assistant node input.",
  "properties": {
    "assistant": {
      "description": "The assistant which will generate the completion.",
      "title": "Assistant",
      "type": "string"
    },
    "prompt": {
      "description": "The text prompt to ask the assistant.",
      "title": "Question",
      "type": "string"
    },
    "memory_key": {
      "description": "A memory key that will keep the chat history shared across runs. Leave empty to leave your assistant without memory of previous messages.",
      "title": "Memory Key",
      "type": "string"
    }
  },
  "required": [
    "assistant",
    "prompt"
  ],
  "title": "AskAssistantInput",
  "type": "object"
}

Output Parameters

Name	Type	Description
messages	array	List of messages from the assistant containing the response.

View JSON Schema

{
  "description": "Response from asking an assistant.",
  "properties": {
    "messages": {
      "description": "List of messages from the assistant.",
      "items": {
        "description": "Message from an assistant.",
        "properties": {
          "id": {
            "type": "string"
          },
          "role": {
            "type": "string"
          },
          "content": {
            "items": {},
            "type": "array"
          },
          "created_at": {
            "type": "integer"
          }
        },
        "type": "object"
      },
      "type": "array"
    }
  },
  "title": "AskAssistantResponse",
  "type": "object"
}

How It Works

This node communicates with an OpenAI assistant through the Assistants API. The assistant is a pre-configured AI entity that may have specific instructions, tools (like code interpreter or web search), and files attached to it. When you ask a question, the node sends your prompt to the assistant and retrieves the response. The optional memory_key parameter allows you to maintain conversation history across multiple workflow runs, enabling persistent conversations with the assistant.

Usage Examples

Example 1: Simple Question to Assistant

Input:

assistant: "asst_abc123def456"
prompt: "What is machine learning?"

Output:

messages: [
  {
    "id": "msg_123",
    "role": "assistant",
    "content": [
      {
        "type": "text",
        "text": {
          "value": "Machine learning is a subset of artificial intelligence..."
        }
      }
    ],
    "created_at": 1699564800
  }
]

Example 2: Question with Memory

Input:

assistant: "asst_coding_helper"
prompt: "Now implement the algorithm in Python"
memory_key: "user_123_conversation"

Output:

messages: [
  {
    "id": "msg_456",
    "role": "assistant",
    "content": [
      {
        "type": "text",
        "text": {
          "value": "Here's a Python implementation..."
        }
      }
    ],
    "created_at": 1699565000
  }
]

Common Use Cases

• Custom AI Assistants: Leverage specialized assistants configured with specific knowledge and tools

• Persistent Conversations: Maintain context across multiple interactions using memory keys

• Document Analysis: Use assistants with document retrieval to analyze uploaded files

• Code Assistance: Leverage code interpreter capabilities built into assistants

• Multi-turn Conversations: Build workflows that handle back-and-forth dialogue

• Tool Integration: Use assistants that have access to functions and external tools

• Domain-Specific Knowledge: Interact with assistants fine-tuned for specific domains

Error Handling

Error Type	Cause	Solution
Assistant Not Found	Selected assistant ID doesn't exist	Verify the assistant ID and ensure it's accessible with your API key
Authentication Error	Invalid or missing OpenAI API key	Verify your OpenAI connection is properly configured
Invalid Prompt	Prompt is empty or malformed	Ensure your prompt is a non-empty string
Memory Key Error	Issue accessing memory storage	Check memory key format and ensure storage is accessible
Rate Limit Error	Too many requests in a short period	Implement delays between requests or upgrade your OpenAI plan
Timeout Error	Request took too long to process	Try with a simpler prompt or shorter expected responses

Notes

• Assistant Selection: Select from your existing OpenAI assistants. Create new assistants in the OpenAI dashboard if needed.

• Memory Keys: Using a memory_key creates a persistent conversation thread. Leaving it empty creates independent, stateless interactions.

• Response Format: Responses are returned as message objects with content arrays that may contain text, code, or other content types.

• Streaming: For long-running operations, consider implementing streaming or async patterns in your workflow.

• Cost: Assistant API calls may have different pricing than ChatGPT API calls. Check OpenAI's pricing page for current rates.