API Reference

Uploading Images

The io.net Intelligence API supports image inputs for vision-enabled AI models. This allows users to send images as part of their API requests, enabling advanced multimodal AI capabilities.

Supported Models for Image Processing

Model NameCapabilities
meta-llama/Llama-3.2-90B-Vision-InstructMulti-modal vision model supporting image understanding.
Qwen/Qwen2-VL-7B-InstructSupports both text and image-based inputs for AI interactions.

📘

Image uploads are only available via API requests. They are not supported in the Chat UI.

Sending an Image via API Request

The API allows two ways to send an image:

  1. Passing an Image URL (recommended for publicly hosted images)
  2. Sending a Base64 Encoded Image (for local images)
import requests

url = "https://api.intelligence.io.solutions/api/v1/chat/completions"

headers = {
    "Authorization": "Bearer $IOINTELLIGENCE_API_KEY",
    "Content-Type": "application/json"
}

data = {
    "model": "meta-llama/Llama-3.2-90B-Vision-Instruct", 
    "messages": [
        {"role": "system", "content": "You are an AI assistant."},
        {"role": "user", "content": [
            {"type": "text", "text": "What is in this image?"},
            {"type": "image_url", "image_url": {"url": "https://your-image-url.com/image.jpg"}}
        ]}
    ]
}

response = requests.post(url, json=data, headers=headers)
print(response.json())


import requests
import base64

url = "https://api.intelligence.io.solutions/api/v1/chat/completions"

headers = {
    "Authorization": "Bearer $IOINTELLIGENCE_API_KEY",
    "Content-Type": "application/json"
}

image_url = "path_to_your_image.jpg"

image_response = requests.get(image_url)
if image_response.status_code == 200:
    image_data = image_response.content  # Get raw image bytes
else:
    print("Error: Unable to download image")
    exit()

base64_image = base64.b64encode(image_data).decode("utf-8")

data = {
    "model": "meta-llama/Llama-3.2-90B-Vision-Instruct", 
    "messages": [
        {"role": "system", "content": "You are an AI assistant."},
        {"role": "user", "content": [
            {"type": "text", "text": "What is in this image?"},
            {"type": "image", "image": base64_image} 
        ]}
    ]
}

response = requests.post(url, json=data, headers=headers)

try:
    print(response.json())  # Parse JSON response
except requests.exceptions.JSONDecodeError:
    print("Error: Unable to parse response. Raw response:", response.text)

📘

The image URL must be publicly accessible. Private or authentication-required URLs won’t work.

Image Input Requirements

To ensure successful processing, images must meet these requirements:

RequirementDetails
FormatJPEG, PNG, WEBP, or GIF (static)
Max File Size20MB
ResolutionAt least 512x512 pixels recommended
Max Dimensions4096×4096 pixels
AccessibilityIf using a URL, ensure it is publicly accessible
Multi-Image SupportUp to 10 images per request

Best Practices for Image Uploads

  • Optimize File Size: While the max limit is 20MB, smaller files (1-5MB) ensure faster processing.
  • Use Clear Images: Avoid blurry or low-resolution images for better AI analysis.
  • Ensure Public URLs: If passing a URL, test it in a browser to confirm it's accessible.

Expected API Response

Upon successful submission, the API will return a structured response with AI-generated insights based on the image.

Example Response:

{
    "id": "chatcmpl-abc123",
    "object": "chat.completion",
    "created": 1710456789,
    "model": "meta-llama/Llama-3.2-90B-Vision-Instruct",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "This is an image of a cat sitting on a table."
            },
            "finish_reason": "stop"
        }
    ],
    "usage": {
        "prompt_tokens": 120,
        "completion_tokens": 20,
        "total_tokens": 140
    }
}

Common Issues & Troubleshooting

IssuePossible CauseSolution
"An image? I'm in text format, so I can't see it..."Model does not support image inputEnsure you are using one of the supported vision models.
"Invalid image format"Image not encoded properlyConvert image to base64 before sending.
"Unauthorized"API key is missing or incorrectCheck that your API key is valid and correctly formatted.