LangChain is a framework for developing applications powered by language models. Z.AI’s integration with LangChain enables you to:
Use LangChain’s chain calling functionality
Build intelligent agents and tool calling
Implement complex conversation memory management
Core Advantages
Framework Ecosystem Access to LangChain’s rich ecosystem and toolchain
Rapid Development Quickly build complex AI applications using pre-built components
Modular Design Flexibly combine different components to meet various needs
Community Support Enjoy active open source community and rich resources
Environment Requirements
Python Version Python 3.8 or higher
LangChain Version langchain_community version 0.0.32 or higher
Please ensure langchain_community version is 0.0.32 or higher for optimal compatibility and feature support.
Install Dependencies Basic Installation # Install LangChain and related dependencies pip install langchain langchainhub httpx_sse # Install OpenAI compatible package pip install langchain-openai Complete Installation # Install all related packages at once pip install langchain langchain-openai langchainhub httpx_sse # Verify installation python -c "import langchain; print(langchain.__version__)" Quick Start Get API Key
Access Z.AI Open Platform , Register or Login.
Create an API Key in the API Keys management page.
Copy your API Key for use.
It is recommended to set the API Key as an environment variable: export ZAI_API_KEY=your-api-key
Basic Configuration import os from langchain_openai import ChatOpenAI # Create Z.AI LLM instance llm = ChatOpenAI( temperature = 0.6 , model = "glm-4.6" , openai_api_key = "your-Z.AI-api-key" , openai_api_base = "https://api.z.ai/api/paas/v4/" ) # Or use environment variables llm = ChatOpenAI( temperature = 0.6 , model = "glm-4.6" , openai_api_key = os.getenv( "ZAI_API_KEY" ), openai_api_base = "https://api.z.ai/api/paas/v4/" ) Basic Usage Examples Simple Conversation from langchain_openai import ChatOpenAI from langchain.schema import HumanMessage, SystemMessage # Create LLM instance llm = ChatOpenAI( temperature = 0.7 , model = "glm-4.6" , openai_api_key = "your-Z.AI-api-key" , openai_api_base = "https://api.z.ai/api/paas/v4/" ) # Create messages messages = [ SystemMessage( content = "You are a helpful AI assistant" ), HumanMessage( content = "Please introduce the development history of artificial intelligence" ) ] # Call the model response = llm(messages) print (response.content) Using Prompt Templates from langchain.prompts import ChatPromptTemplate from langchain_openai import ChatOpenAI # Create LLM llm = ChatOpenAI( model = "glm-4.6" , openai_api_key = "your-Z.AI-api-key" , openai_api_base = "https://api.z.ai/api/paas/v4/" ) # Create prompt template prompt = ChatPromptTemplate.from_messages([ ( "system" , "You are a professional {domain} expert" ), ( "human" , "Please explain the concept and applications of {topic} " ) ]) # Create chain chain = prompt | llm # Invoke chain response = chain.invoke({ "domain" : "machine learning" , "topic" : "deep learning" }) print (response.content) Conversation Memory Management from langchain_openai import ChatOpenAI from langchain.prompts import ( ChatPromptTemplate, MessagesPlaceholder, SystemMessagePromptTemplate, HumanMessagePromptTemplate, ) from langchain.chains import LLMChain from langchain.memory import ConversationBufferMemory # Create LLM llm = ChatOpenAI( temperature = 0.6 , model = "glm-4.6" , openai_api_key = "your-Z.AI-api-key" , openai_api_base = "https://api.z.ai/api/paas/v4/" ) # Create prompt template prompt = ChatPromptTemplate( messages = [ SystemMessagePromptTemplate.from_template( "You are a nice chatbot having a conversation with a human." ), MessagesPlaceholder( variable_name = "chat_history" ), HumanMessagePromptTemplate.from_template( " {question} " ) ] ) # Create memory memory = ConversationBufferMemory( memory_key = "chat_history" , return_messages = True ) # Create conversation chain conversation = LLMChain( llm = llm, prompt = prompt, verbose = True , memory = memory ) # Have conversations response1 = conversation.invoke({ "question" : "tell me a joke" }) print ( "AI:" , response1[ 'text' ]) response2 = conversation.invoke({ "question" : "tell me another one" }) print ( "AI:" , response2[ 'text' ]) Advanced Features Intelligent Agent import os from langchain import hub from langchain.agents import AgentExecutor, create_react_agent from langchain_community.tools.tavily_search import TavilySearchResults from langchain_openai import ChatOpenAI # Set search tool API key os.environ[ "TAVILY_API_KEY" ] = "your-tavily-api-key" # Create LLM llm = ChatOpenAI( model = "glm-4.6" , openai_api_key = "your-Z.AI-api-key" , openai_api_base = "https://api.z.ai/api/paas/v4/" ) # Create tools tools = [TavilySearchResults( max_results = 2 )] # Get prompt template prompt = hub.pull( "hwchase17/react" ) # Create agent agent = create_react_agent(llm, tools, prompt) agent_executor = AgentExecutor( agent = agent, tools = tools, verbose = True ) # Execute task result = agent_executor.invoke({ "input" : "what is LangChain?" }) print (result[ 'output' ]) Streaming Output from langchain_openai import ChatOpenAI from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler from langchain.schema import HumanMessage # Create LLM with streaming output llm = ChatOpenAI( model = "glm-4.6" , openai_api_key = "your-Z.AI-api-key" , openai_api_base = "https://api.z.ai/api/paas/v4/" , streaming = True , callbacks = [StreamingStdOutCallbackHandler()] ) # Send message (output will be displayed in real-time streaming) response = llm([HumanMessage( content = "Write a poem about spring" )]) Best Practices
Performance Optimization
Enable LangChain caching mechanism
Use batch processing to reduce API calls
Set reasonable max_tokens limits
Use async processing for better concurrency
Error Handling
Implement retry mechanisms and exponential backoff
Set reasonable timeout values
Log detailed error information
Provide fallback solutions
Memory Management
Use ConversationBufferWindowMemory to limit history length
Regularly clean unnecessary conversation history
Monitor memory usage
Implement conversation summarization
Security
Use environment variables to store API keys
Implement input validation and filtering
Monitor API usage and costs
Rotate API keys regularly
Getting Help LangChain is a rapidly evolving framework. It is recommended to update to the latest version regularly for optimal functionality and performance. Meanwhile, Z.AI will continue to optimize integration with LangChain to ensure the best compatibility and user experience.
