How to Use OpenAI's Vision API to Caption Images Automatically

For local images, encode them as base64:

import base64
from pathlib import Path

def caption_local_image(file_path, style="descriptive"):
    image_data = Path(file_path).read_bytes()
    base64_image = base64.b64encode(image_data).decode("utf-8")

    # Detect MIME type
    ext = Path(file_path).suffix.lower()
    mime_types = {".jpg": "image/jpeg", ".jpeg": "image/jpeg", ".png": "image/png", ".webp": "image/webp", ".gif": "image/gif"}
    mime = mime_types.get(ext, "image/jpeg")

    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[
            {
                "role": "user",
                "content": [
                    {"type": "text", "text": "Describe this image in one detailed sentence."},
                    {"type": "image_url", "image_url": {"url": f"data:{mime};base64,{base64_image}"}},
                ],
            }
        ],
        max_tokens=150,
    )
    return response.choices[0].message.content

print(caption_local_image("./photos/beach.jpg"))

Process an entire folder of images and save captions to a JSON file:

import json
import os
from pathlib import Path

def batch_caption(folder, output_file="captions.json", style="descriptive"):
    supported = {".jpg", ".jpeg", ".png", ".webp", ".gif"}
    results = []

    images = [f for f in Path(folder).iterdir() if f.suffix.lower() in supported]
    print(f"Found {len(images)} images to caption")

    for i, img_path in enumerate(images, 1):
        print(f"  [{i}/{len(images)}] {img_path.name}...", end=" ")
        try:
            caption = caption_local_image(str(img_path), style)
            results.append({"file": img_path.name, "caption": caption})
            print("Done")
        except Exception as e:
            results.append({"file": img_path.name, "error": str(e)})
            print(f"Error: {e}")

    with open(output_file, "w") as f:
        json.dump(results, f, indent=2)

    print(f"\nSaved {len(results)} captions to {output_file}")

batch_caption("./product-images")

This generates a clean JSON file you can use for alt text, image SEO, or content management.

Important considerations for production use:

To reduce costs:

• Use detail: "low" in the image_url object for simple captioning tasks
• Resize large images before sending (1024px width is sufficient)
• Cache results to avoid re-captioning the same images

GPT-4o produces highly accurate and detailed descriptions. For accessibility alt text, the quality is typically better than human-written captions because the model describes what it actually sees.

Yes, add the target language to your prompt: "Describe this image in one sentence in Spanish."

Reformat offers a free AI Image Caption tool that generates captions without needing an API key. It is great for quick, one-off captioning tasks.