Guides
Vision
Claude, GPT and Gemini models can read images: describe what's in them, extract text, and analyze charts and UI screenshots. Put the image and your text in the same user message.
Two ways to send an image
- Image URL: a publicly reachable
https://address. - Base64: encode a local image and put it in the request. Good for private images.
OpenAI-compatible format
vision.py
from openai import OpenAI
client = OpenAI(api_key="sk-...", base_url="https://api.flyrouter.ai/v1")
resp = client.chat.completions.create(
model="gpt-6.1-sol",
messages=[{
"role": "user",
"content": [
{"type": "text", "text": "What's in this image?"},
{"type": "image_url", "image_url": {"url": "https://example.com/photo.jpg"}},
],
}],
)
print(resp.choices[0].message.content)Anthropic native format
vision_claude.py
import base64
from anthropic import Anthropic
client = Anthropic(api_key="sk-...", base_url="https://api.flyrouter.ai")
with open("chart.png", "rb") as f:
data = base64.b64encode(f.read()).decode()
msg = client.messages.create(
model="claude-sonnet-5-5",
max_tokens=2048,
messages=[{
"role": "user",
"content": [
{"type": "image", "source": {"type": "base64", "media_type": "image/png", "data": data}},
{"type": "text", "text": "Summarize the main trend in this chart"},
],
}],
)
print(msg.content[0].text)To use an image URL instead, set source to {"type": "url", "url": "https://..."}.
Things to know
- Common formats such as JPEG, PNG, GIF and WebP are supported.
- Images are billed as input tokens based on their size, so bigger images cost more. Resizing images to the resolution you actually need before sending can cut costs noticeably.
- A single message can contain several images for comparison.
- Placing images before the question text usually works better.
FlyRouter