It seems do not support vision request?

[Summer-Shen98]

I have add "Copilot-Vision-Request": "true" to the headers of config.json
but when i use the following code to analyze image it raise: {'error': {'message': 'missing required Copilot-Vision-Request header for vision requests', 'code': ''}}

import base64
from openai import OpenAI

client = OpenAI(
base_url="http://10.40.2.26:8081/v1",
api_key="test-key"
)
with open("1.jpg", "rb") as f:
img_base64 = base64.b64encode(f.read()).decode("utf-8")

response = client.chat.completions.create(
model="gpt-4o",
messages=[
{
"role": "user",
"content": [
{"type": "text", "text": "what is in the picture?"},
{
"type": "image_url",
"image_url": {
"url": f"data:image/jpeg;base64,{img_base64}"
}
}
]
}
]
)

print(response.choices[0].message.content)

[privapps]

Apologies for the delayed response. You are correct that the current implementation does not support vision. To add a header, you would need to modify the proxy code to allow the Copilot-Vision-Request to pass through.

I ran some quick tests and noticed that there are additional areas in the code that would require changes as well.

To be honest, the primary use of this app is for coding. It works seamlessly with nearly all coding CLI tools that supports OpenAI chat completions format, along with a few AI plugins for VS Code such as Cline, Roo Code, and others. Personally, I rarely find the need for vision functionality in my use case.

One thing to keep in mind is that GitHub Copilot has a content length limit of 64k. Even if the vision feature is enabled, this limitation would still restrict its usage.

That said, if you decide to create a patch for this, I’d be happy to review and merge it.

[jett2025]

Does the proxy support model thinking capability (o1/extended thinking)?