LangChain Integration with DeepSeek

This chapter explains how to integrate and use the DeepSeek chat model in LangChain.

Through thelangchain-deepseekextension package, developers can quickly access the large language model services provided by DeepSeek.

DeepSeek models support both the official hosted API and local or third-party inference deployment through platforms such as Ollama, Fireworks, and Together.

In addition to the official API interface, we can also use theCoding Plan/Token Planpackage to directly access mainstream large models such as DeepSeek, Kimi, GLM, Doubao, MiniMax, etc., without purchasing each vendor's APIs separately.

DeepSeek Introduction

DeepSeek is an open-source large language model series that supports capabilities such as chat, reasoning, and code generation.

In LangChain, DeepSeek is mainly called through theChatDeepSeekclass.

Integration Information:

Item Description
Class Name ChatDeepSeek
Package langchain-deepseek
Status Beta (Test Version)
JavaScript Support Yes
Python Support Yes

Model Capability Support:

Feature Supported?
Tool Calling ✅ Supported
Structured Output ✅ Supported
Image Input ❌ Not Supported
Audio Input ❌ Not Supported
Video Input ❌ Not Supported
Token Streaming Output ✅ Supported
Native Async Calling ✅ Supported
Token Usage Statistics ✅ Supported
Logprobs ❌ Not Supported

The DeepSeek API uses an API format compatible with OpenAI/Anthropic. By modifying the configuration, you can use the OpenAI/Anthropic SDK to access the DeepSeek API, or use software compatible with the OpenAI/Anthropic API.

ParameterValue
base_url (OpenAI)https://api.deepseek.com
base_url (Anthropic)https://api.deepseek.com/anthropic
api_keyClick the link to applyAPI key
model*deepseek-v4-flash
deepseek-v4-pro
deepseek-chat(Will be deprecated on 2026/07/24)
deepseek-reasoner(Will be deprecated on 2026/07/24)

Install langchain-deepseek

Before using it, you need to install DeepSeek's LangChain integration package:

pip install -qU langchain-deepseek

Configure DeepSeek API Key

You need to first register a DeepSeek account and create an API Key. If you don't have one yet, you need to go tohttps://platform.deepseek.com/api_keysto create an API key.

It is usually recommended to usepython-dotenvto read the.envconfiguration file in the current directory.

pip install python-dotenv

Create the.envfile in the current project directory:

.env File Configuration:

DEEPSEEK_API_KEY="sk-xxxxxxxxxxxxxxxx"
OPENAI_API_KEY="sk-xxxxxxxxxxxxxxxx"

Python Reads the .env File:

Complete Example

import os

from dotenv import load_dotenv

# Load the .env file in the current directory
load_dotenv()

# Get API Key
api_key = os.getenv("DEEPSEEK_API_KEY")

print(api_key)

Specify the .env File Path:If the .env file is not in the current directory, you can manually specify the path:

Example

from dotenv import load_dotenv

load_dotenv(dotenv_path="./config/.env")

Configure LangSmith (Optional)

If you need to enable LangChain call chain tracing and debugging, you can configure the LangSmith API Key:

Example

os.environ["LANGSMITH_TRACING"] = "true"

os.environ["LANGSMITH_API_KEY"] = getpass.getpass(
    "Enter your LangSmith API key: "
)

Create ChatDeepSeek Model

After installation, you can useChatDeepSeekto initialize the model:

Example

import os

from dotenv import load_dotenv
from langchain_deepseek import ChatDeepSeek

# Load .env
load_dotenv()

# Get API KEY
api_key = os.getenv("DEEPSEEK_API_KEY")

# Create model
llm = ChatDeepSeek(
    api_key=api_key,
    model="deepseek-v4-flash",
    temperature=0,
    max_tokens=None,
    timeout=None,
    max_retries=2
)

# Call model
response = llm.invoke("Hello, please introduce LangChain")

print(response.content)

Parameter Description:

Parameter Description
api_key Set your applied API key, e.g., "sk-xxx"
model Specify the model name, e.g., deepseek-v4-flash
temperature Controls randomness; the lower the value, the more stable the result
max_tokens Limit the maximum number of generated tokens
timeout Request timeout
max_retries Maximum number of retries after failure

You can also use theinit_chat_model()function to call:

Example

import os

from dotenv import load_dotenv

# Load the .env file in the current directory
load_dotenv()
from langchain.chat_models import init_chat_model

# Specified model, returns a fixed model
model = init_chat_model("deepseek:deepseek-v4-flash", temperature=0.7)
response = model.invoke("Introduce the Python Tutorial")
print(response.content)

Call DeepSeek Model

The following example demonstrates how to send chat messages to the DeepSeek model:

Note: In my test here, I wrote the API Key in the test file:

llm = ChatDeepSeek(
    api_key="sk-xxx",  # 设置你的 DeepSeek API Key
    model="deepseek-v4-flash"
)

In actual production environments, please set it in the.envfile.

Example

from langchain_deepseek import ChatDeepSeek

llm = ChatDeepSeek(
    api_key="sk-xxx",    # Set your DeepSeek API Key
    model="deepseek-v4-flash",
    temperature=0,
    max_tokens=None,
    timeout=None,
    max_retries=2
)

messages = [
    (
        "system",
        "You are a helpful assistant that translates English to Chinese."
    ),
    (
        "human",
        "I love programming."
    ),
]

ai_msg = llm.invoke(messages)

print(ai_msg.content)

The above code translates English into Chinese. After running, the model will return the translated result:

I like programming.

Supported DeepSeek Models:

Model Purpose Features
deepseek-v4-flash General-purpose chat model Supports Tool Calling and structured output
deepseek-v4-pro Reasoning model (deepseek-v4-pro) Stronger reasoning capabilities, but does not support Tool Calling

LangChain + DeepSeek Complete Test Example

The following demonstrates a complete runnable LangChain + DeepSeek example.

Create Test File

Create the test.py file:

Complete Example

from langchain_deepseek import ChatDeepSeek
from langchain_core.messages import HumanMessage, SystemMessage

# =========================
# Configure DeepSeek API Key
# =========================

apiKey = "sk-xxx" # Set your DeepSeek API Key

# =========================
# Create DeepSeek model
# =========================

llm = ChatDeepSeek(
    api_key=apiKey,
    model="deepseek-v4-flash",
    temperature=0.7,
    max_tokens=1024,
    timeout=60,
    max_retries=3
)

# =========================
# Construct chat messages
# =========================

messages = [
    SystemMessage(
        content="You are a professional Python teacher."
    ),

    HumanMessage(
        content="Please explain what LangChain is, and provide a simple example."
    )
]

# =========================
# Call model
# =========================

response = llm.invoke(messages)

# =========================
# Output result
# =========================

print("AI Reply:")
print(response.content)

Execute in the terminal:

python test.py

The output is as follows:

Code Explanation:

Code Purpose
ChatDeepSeek LangChain's DeepSeek chat model class
SystemMessage System prompt, used to set the AI's identity
HumanMessage User input message
llm.invoke() Call the DeepSeek model
response.content Get the AI's returned content

Use the deepseek-v4-pro Reasoning Model

LangChain supports DeepSeek streaming output, requiring the deepseek-v4-pro inference model:

Example

from langchain_deepseek import ChatDeepSeek
from langchain_core.messages import HumanMessage, SystemMessage

# =========================
# Configure DeepSeek API Key
# =========================

apiKey = "sk-xxx" # Set your DeepSeek API Key

# =========================
# Create DeepSeek model
# =========================

llm = ChatDeepSeek(
    api_key=apiKey,
    model="deepseek-v4-pro",
    temperature=0.7,
    max_tokens=1024,
    timeout=60,
    max_retries=3
)

for chunk in llm.stream("Please introduce Python"):
    print(chunk.content, end="", flush=True)

Use PromptTemplate

Dynamically generate Prompt with PromptTemplate:

Example

from langchain_deepseek import ChatDeepSeek
from langchain_core.prompts import ChatPromptTemplate

# =========================
# Configure DeepSeek API Key
# =========================

apiKey = "sk-xxx" # Set your DeepSeek API Key

# =========================
# Create DeepSeek model
# =========================

llm = ChatDeepSeek(
    api_key=apiKey,
    model="deepseek-v4-pro",
    temperature=0.7,
    max_tokens=1024,
    timeout=60,
    max_retries=3
)

prompt = ChatPromptTemplate.from_template(
    "Please explain in detail: {topic}"
)

chain = prompt | llm

response = chain.invoke({
    "topic": "Transformer"
})

print(response.content)

Common Errors

Error Cause Solution
401 Unauthorized API Key error Check DEEPSEEK_API_KEY
ModuleNotFoundError Dependencies not installed Re-run pip install
Rate Limit Request frequency too high Reduce request frequency
Timeout Error Request timeout Increase timeout parameter

Recommended Project Structure

Project Structure

project/
│
├── app.py
├── requirements.txt
├── .env
├── prompts/
├── data/
└── vector_db/

requirements.txt Example

Example

langchain
langchain-deepseek
python-dotenv

Reference Documentation

Other extensions