Claude Vision API in Python

Analyze images with Claude in Python. Pass base64 images or URLs, extract text, describe scenes, and process PDFs with the Anthropic vision API.

Claude's vision capability lets you analyze images inline in the messages array — no separate endpoint required.

Analyze an image from a URL

import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "source": {
                        "type": "url",
                        "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/4/47/PNG_transparency_demonstration_1.png/280px-PNG_transparency_demonstration_1.png"
                    }
                },
                {"type": "text", "text": "What do you see in this image? Be specific."}
            ]
        }
    ]
)
print(response.content[0].text)

Analyze a local image (base64)

import anthropic
import base64
from pathlib import Path

client = anthropic.Anthropic()

def analyze_local_image(image_path: str, prompt: str) -> str:
    image_data = base64.standard_b64encode(Path(image_path).read_bytes()).decode()

    # Detect media type from extension
    ext = Path(image_path).suffix.lower()
    media_type = {"jpg": "image/jpeg", ".jpeg": "image/jpeg",
                  ".png": "image/png", ".gif": "image/gif",
                  ".webp": "image/webp"}.get(ext, "image/jpeg")

    response = client.messages.create(
        model="claude-sonnet-4-6",
        max_tokens=1024,
        messages=[
            {
                "role": "user",
                "content": [
                    {
                        "type": "image",
                        "source": {"type": "base64", "media_type": media_type, "data": image_data}
                    },
                    {"type": "text", "text": prompt}
                ]
            }
        ]
    )
    return response.content[0].text

# Usage
result = analyze_local_image("screenshot.png", "Extract all text visible in this screenshot.")
print(result)

Multiple images in one request

response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "image", "source": {"type": "url", "url": "https://example.com/before.jpg"}},
                {"type": "text", "text": "Before:"},
                {"type": "image", "source": {"type": "url", "url": "https://example.com/after.jpg"}},
                {"type": "text", "text": "After: What changed between these two images?"}
            ]
        }
    ]
)

Use the Cost Calculator to estimate token costs for image workloads. See the Python quickstart for non-vision API basics.

Frequently asked questions

Free tools

More examples