LangChain integration
Use DeepSeek in LangChain Through Tokuse and ChatOpenAI
LangChain can use an OpenAI-compatible chat model when the provider follows the chat completions contract. ChatOpenAI keeps the integration small and lets you change the model alias later.
Settings
- Class
- ChatOpenAI
- Base URL
- https://tokuse.com/v1
- Model
- deepseek-flash
Setup steps
- Install langchain-openai and the OpenAI SDK.
- Keep TOKUSE_API_KEY in the server environment.
- Configure ChatOpenAI with base_url https://tokuse.com/v1.
- Select a model currently available in your Tokuse account.
- Add timeout, retry, and evaluation logic before production rollout.
Configuration
import os
from langchain_openai import ChatOpenAI
llm = ChatOpenAI(
model="deepseek-flash",
api_key=os.environ["TOKUSE_API_KEY"],
base_url="https://tokuse.com/v1",
timeout=30,
max_retries=2,
)
result = llm.invoke("Return a JSON object with a one-line release summary.")
print(result.content)Before production use
- Keep the Tokuse key on the server or in the tool's secret store.
- Use the base URL without appending /chat/completions.
- Verify the model alias in the current Tokuse catalog.
- Test errors, timeout behavior, and cost with a small workload first.