Langchain Usage Instructions

Interface Information

Item	Value
Base URL (for LangChain)	https://api-ai.gitcode.com/v1
Full Chat Path	POST https://api-ai.gitcode.com/v1/chat/completions
Model ID	deepseek-ai/DeepSeek-R1
Protocol	JSON compatible with OpenAI Chat Completions

ChatOpenAI will automatically append /chat/completions to base_url, do not set base_url as a full URL including /chat/completions.

Authentication

Usually it is Bearer Token (please refer to the platform's actual instructions):

Authorization: Bearer <AtomGit Access Token>
Content-Type: application/json

Request Body (same as OpenAI, example)

{
  "model": "deepseek-ai/DeepSeek-R1",
  "messages": [
    { "role": "system", "content": "You are a helpful assistant." },
    { "role": "user", "content": "Introduce LangChain in one sentence." }
  ],
  "temperature": 0.7,
  "max_tokens": 1024,
  "stream": true
}

Optional fields (if supported by the server): top_p, stop, presence_penalty, frequency_penalty, etc.

---

Environment Setup

pip install langchain-openai

---

Example One: Streaming Chat

import os
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage, SystemMessage

llm = ChatOpenAI(
    model="deepseek-ai/DeepSeek-R1",
    base_url="https://api-ai.gitcode.com/v1",
    api_key=os.environ["ATOMGit_AI_API_KEY"],  # AtomGit Access Token (API key)
    temperature=0.7,
    streaming=True,
)

messages = [
    SystemMessage(content="You are a helpful assistant."),
    HumanMessage(content="Introduce LangChain in one sentence."),
]

for chunk in llm.stream(messages):
    if chunk.content:
        print(chunk.content, end="", flush=True)
print()

---

Notes

1. base_url: Ends with /v1, do not repeat /chat/completions.
2. model: Must match the platform's model: deepseek-ai/DeepSeek-R1.
3. Access Token: Do not hardcode the AtomGit access token into the repository, use environment variables (such as ATOMGit_AI_API_KEY) or include it in Authorization: Bearer <AtomGit Access Token>.