copilot-client 1.0.0

Lightweight GitHub Copilot client with device login

Copilot Python Client

Synchronous GitHub Copilot client and CLI with device flow auth, token exchange, model management, and streaming chat.

Package/CLI version: 1.0.0

Installation

• Python 3.10+
• From PyPI (recommended):

  pip install copilot-client
  

• From GitHub:

  pip install "git+https://github.com/AlpMeteSenel/copilot-python-client.git"
  

CLI (copilot-client)

The CLI is installed as copilot-client (same version as the package). Credentials live in ~/.copilot_client/config.json and are reused across commands.

• Authenticate via device flow or paste a token: copilot-client auth
• List available models: copilot-client models
• Interactive chat: copilot-client chat --model gpt-5-mini --system "Be concise"

Quickstart (programmatic)

import os
from copilot_client import CopilotClient

# Use an existing Copilot access token or fetch one via device flow (see below)
client = CopilotClient(copilot_access_token=os.environ["COPILOT_TOKEN"])
print(client.get_user())
print(client.chat([{ "role": "user", "content": "Say hi" }]))

Auth options

• Device flow:

  code = CopilotClient.start_device_flow()
  print(f"Open {code.verification_uri} and enter {code.user_code}")
  github_token = CopilotClient.poll_device_flow(code.device_code, poll_interval=code.interval)
  client = CopilotClient(copilot_access_token=github_token)
  print("Copilot token:", client.copilot_token)
  

• Existing Copilot access token:

  client = CopilotClient(copilot_access_token=os.environ["COPILOT_TOKEN"])
  

End-to-end walkthrough

from pathlib import Path
import base64
from copilot_client import CopilotClient

# Start device flow and exchange for a Copilot token
code = CopilotClient.start_device_flow()
print(f"Open {code.verification_uri} and enter {code.user_code}")
token = CopilotClient.poll_device_flow(code.device_code, poll_interval=code.interval)
client = CopilotClient(copilot_access_token=token)

print("copilot_token", client.copilot_token)
print("user", client.get_user())

models = client.list_models()
print([[m["id"], (m.get("policy", {}).get("state", "no policy"))] for m in models])

if models:
    try:
        print("enable_model", client.enable_model(models[0]["id"]))
    except Exception as exc:
        print("enable_model skipped", exc)

reply = client.chat([{"role": "user", "content": "Summarize GitHub in 2 bullets"}], system_message="Be concise")
print("chat", reply)

print("chat_stream", end=" ")
for chunk in client.chat_stream([{ "role": "user", "content": "Name two Python async primitives"}], model="copilot-nes-xtab"):
    if chunk["type"] == "text-delta":
        print(chunk["text"], end="")
print()

image_path = Path("husky.jpg")  # replace with your image path
image_base64 = base64.b64encode(image_path.read_bytes()).decode("ascii")
vision_message = [{
    "role": "user",
    "content": [
        {"type": "text", "text": "What is on this image?"},
        {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{image_base64}", "detail": "high"}},
    ],
}]

for chunk in client.chat_stream(vision_message, model="gpt-5-mini", vision=True):
    if chunk["type"] == "text-delta":
        print(chunk["text"], end="")
print()

Core usage

• Models: list with client.list_models(); enable those needing consent via client.enable_model(model_id).
• Chat (non-streaming): client.chat([...], model="gpt-5-mini", system_message="Be concise", max_output_tokens=200).
• Chat (streaming): iterate over client.chat_stream([...], model="copilot-nes-xtab") and print chunk["text"] when chunk["type"] == "text-delta".
• User info: client.get_user().

enable_model only activates models that require explicit consent; many will already be enabled.

Errors

All methods raise CopilotError on non-2xx responses. Auth issues raise CopilotAuthError; rate limits raise CopilotRateLimitError.

Tips

• chat_stream(..., vision=True) forwards the Copilot vision header when prompts include images.
• Copilot access tokens are exchanged automatically for Copilot bearer tokens and refreshed every 30 minutes.

Credits

Built by Alp Mete Senel. Contributions and issues are welcome.