Get Answer

Get an AI-generated response to user inputs with support for multi-modal content including text, images, and files.

Endpoint

PUT /api/v1/assistant/get-answer

Authentication

Requires valid session authentication via middleware.

Request Format

The request must be sent as multipart/form-data to support file uploads.

Parameters

Parameter	Type	Required	Description
inputs	Array	Yes	Array of conversation inputs
toolKitSettings	Object	Yes	AI model configuration
files	File[]	No	Uploaded files (images, documents)

Input Object Structure

{
  "role": "user", // "user" | "assistant" | "system"
  "value": "Your question or message",
  "imageUrls": ["url1", "url2"], // Optional image URLs
  "files": [
    {
      "name": "document.pdf",
      "id": "file-id",
      "type": "application/pdf"
    }
  ]
}

ToolKit Settings

{
  "model": "gpt-4o", // AI model to use
  "temperature": 0.7, // 0.0 - 1.0, creativity level
  "chatLang": "english", // "english" | "hebrew" | "none"
  "isMultiple": false, // Multiple responses
  "isLongAnswer": false, // Extended response format
  "isTypingEffect": true, // Typing animation
  "question_choice": ["Vectors"], // ["Vectors", "Q&A"]
  "embeddingModel": "text-embedding-ada-002" // For document search
}

Example Request

const formData = new FormData();

// Add conversation inputs
formData.append(
    'inputs',
    JSON.stringify([
        {
            role: 'user',
            value: 'Explain quantum computing in simple terms',
        },
    ])
);

// Add toolkit settings
formData.append(
    'toolKitSettings',
    JSON.stringify({
        model: 'gpt-4o',
        temperature: 0.7,
        chatLang: 'english',
        isMultiple: false,
        isLongAnswer: false,
        isTypingEffect: true,
    })
);

// Add files if any
formData.append('files', fileInput.files[0]);

const response = await fetch('/api/v1/assistant/get-answer', {
    method: 'PUT',
    body: formData,
});

Response

Success Response (200)

{
    "QuestionInEnglish": "Explain quantum computing in simple terms",
    "AnsFromAi": "Quantum computing is a revolutionary approach to computation that...",
    "categoryText": "",
    "sentimentText": ""
}

Response Fields

Field	Type	Description
QuestionInEnglish	string	The user's question translated to English
AnsFromAi	string	The AI-generated response
categoryText	string	Category classification (optional)
sentimentText	string	Sentiment analysis (optional)

Error Responses

400 Bad Request

{
    "error": "Invalid payload structure"
}

429 Too Many Requests

{
    "error": "Input is too long for requested model."
}

500 Internal Server Error

{
    "error": "AI service temporarily unavailable"
}

Supported Models

• OpenAI: gpt-4, gpt-4o, gpt-4o-mini
• Anthropic: claude-3-opus, claude-3-sonnet, claude-3-haiku
• Google: gemini-1.5-pro, gemini-1.5-flash, gemini-2.0-flash
• Other: jamba-large, mistral-large-latest

File Support

Supported file types:

• Images: PNG, JPEG, GIF, WebP
• Documents: PDF, DOCX, TXT, MD
• Data: CSV, JSON, XML

Maximum file size: 50MB per request

Notes

• The endpoint automatically handles language detection and translation
• Image processing requires models that support vision capabilities
• Large files are processed asynchronously in the background
• Session context is maintained for conversation continuity