# 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. |

<details>

<summary>View JSON Schema</summary>

```json
{
  "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"
}
```

</details>

## Output Parameters

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

<details>

<summary>View JSON Schema</summary>

```json
{
  "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"
}
```

</details>

## 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.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.agenticflow.ai/reference/nodes/openai_ask_assistant.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.