Many models can read images: screenshots, photos, receipts, charts, handwritten notes. You send the image inside a normal chat message.
Image from a URL
r = client.chat.completions.create(
model="claude-sonnet-4-6",
messages=[{
"role": "user",
"content": [
{"type": "text", "text": "What's wrong with this UI? Give 3 concrete fixes."},
{"type": "image_url", "image_url": {"url": "https://example.com/screenshot.png"}},
],
}],
)Image from a local file (base64)
import base64
with open("receipt.jpg", "rb") as f:
b64 = base64.b64encode(f.read()).decode()
r = client.chat.completions.create(
model="gpt-5.6-luna",
messages=[{"role": "user", "content": [
{"type": "text", "text": "Extract merchant, date and total as JSON."},
{"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{b64}"}},
]}],
)Good uses
- Receipts and invoices: extract fields into JSON (see extract data from invoices).
- UI review: paste a screenshot and ask for accessibility or layout problems.
- Charts: “summarise the trend in this chart”.
- Bug reports: a screenshot of an error dialog is often clearer than a description.
Cost tips
Images are converted into input tokens, and large images cost more. Resize screenshots to the size that's still readable (around 1,000–1,500 px on the long side is usually plenty) before sending.
Check a model's capabilities on the model page: not every model accepts images.
Claude, GPT, Gemini, DeepSeek, image and video models through one OpenAI- and Anthropic-compatible endpoint. Prepaid, pay only for what you use.