LangChain & LangGraph Examples

A collection of provider-agnostic examples demonstrating LangChain and LangGraph patterns.

Setup

Follow these steps to get the examples running locally.

Prerequisites

• Python 3.9 or later
• git (to clone the repository)
• pip (included with Python)
• Optional: Ollama for local models

Step 1: Clone the repository

git clone https://github.com/kenwith/langchain_examples.git
cd langchain_examples

Step 2: Create a virtual environment

python -m venv .venv

Step 3: Activate the virtual environment

# On macOS/Linux
source .venv/bin/activate

# On Windows
.venv\Scripts\activate

Step 4: Install dependencies

pip install -r requirements.txt

Step 5: Configure environment variables

cp .env.example .env

Open .env in a text editor and add the API keys for the providers you plan to use. See API Key Configuration below.

Step 6: Verify setup

Run the first example:

python examples/01_basic_chains.py

If everything is configured correctly, you should see a generated response from your chosen model.

Project Structure

langchain_examples/
├── examples/          # Runnable example scripts
│   ├── 01_basic_chains.py
│   ├── 02_rag.py
│   ├── 03_tools_agents.py
│   ├── 04_langgraph_workflows.py
│   ├── 05_streaming.py
│   ├── 06_evaluation.py
│   ├── 07_memory.py
│   ├── 08_batch_inference.py
│   ├── 09_parallel_tool_calls.py
│   ├── 10_async.py
│   └── 11_parallel_tool_calls.py
├── .env.example       # Template for environment variables
├── requirements.txt   # Python dependencies
└── README.md          # This file

Model Configuration

All examples use init_chat_model for provider-agnostic model selection. Configure via environment variable or .env:

# Environment variable (takes precedence)
export LANGCHAIN_MODEL="anthropic/claude-3-5-sonnet"
export LANGCHAIN_MODEL="openai/gpt-4o"
export LANGCHAIN_MODEL="google/gemini-1.5-pro"
export LANGCHAIN_MODEL="groq/llama-3.1-70b-versatile"
export LANGCHAIN_MODEL="ollama/llama3.1"

# Or in .env file
LANGCHAIN_MODEL=anthropic/claude-3-5-sonnet
# Optional: override base URL for OpenAI-compatible APIs
LANGCHAIN_API_BASE=https://api.example.com/v1

See LangChain init_chat_model providers for full provider list and configuration options.

API Key Configuration

The examples load API keys from your environment or .env file. Add only the keys for the providers you intend to use.

Provider	Environment variable	Model format	Notes
Anthropic	ANTHROPIC_API_KEY	anthropic/claude-3-5-sonnet	Key starts with sk-ant-
OpenAI	OPENAI_API_KEY	openai/gpt-4o	Key starts with sk-
Google	GOOGLE_API_KEY	google/gemini-1.5-pro	Enable Generative Language API in Google Cloud Console
Groq	GROQ_API_KEY	groq/llama-3.1-70b-versatile	Key starts with gsk_
Ollama	No key required	ollama/llama3.1	Requires local server at http://localhost:11434

Example .env entries:

# Choose your default model
LANGCHAIN_MODEL=anthropic/claude-3-5-sonnet

# Provider API keys (only add the ones you need)
ANTHROPIC_API_KEY=your-anthropic-api-key-here
OPENAI_API_KEY=your-openai-api-key-here
GOOGLE_API_KEY=your-google-api-key-here
GROQ_API_KEY=your-groq-api-key-here

Never commit real API keys. The .env file is listed in .gitignore.

Examples

Category	File	Description
Basic Chains	examples/01_basic_chains.py	LLM + prompt + output parser
RAG	examples/02_rag.py	Retrieval-augmented generation
Tools/Agents	examples/03_tools_agents.py	Function calling, ReAct agent
LangGraph Workflows	examples/04_langgraph_workflows.py	Stateful multi-step workflows
Streaming	examples/05_streaming.py	Streaming responses
Evaluation	examples/06_evaluation.py	Testing and evaluation patterns
Memory	examples/07_memory.py	Conversation history management
Batch Inference	examples/08_batch_inference.py	Process multiple inputs efficiently with batch API calls
Tools/Agents	examples/09_parallel_tool_calls.py	Execute multiple tool calls in parallel with a single model response
Async	examples/10_async.py	Async/await patterns for concurrent model calls
Tools/Agents	examples/11_parallel_tool_calls.py	Parallel tool calls with async execution

Usage Examples

01_basic_chains.py

Demonstrates a basic LLM chain with a prompt template and output parser.

python examples/01_basic_chains.py

To override the model for this run:

LANGCHAIN_MODEL=openai/gpt-4o python examples/01_basic_chains.py

Expected output: a parsed, structured response generated by the model based on the prompt.

02_rag.py

Shows retrieval-augmented generation using a vector store and document loaders.

python examples/02_rag.py

If the script uses external documents, make sure the data files are in the expected location (check the script for DATA_PATH or similar). You can also modify the script to load your own documents.

03_tools_agents.py

Demonstrates function calling and a ReAct agent with tools.

python examples/03_tools_agents.py

The agent will use the defined tools to answer a question. If the script requires network access for tools like search, ensure your environment allows it.

04_langgraph_workflows.py

Builds a stateful multi-step workflow with LangGraph.

python examples/04_langgraph_workflows.py

This example shows how to define nodes, edges, and state transitions in a LangGraph workflow. Review the output to see each step execute in order.

05_streaming.py

Streams model responses token-by-token.

python examples/05_streaming.py

Some providers require explicit streaming support. If streaming does not work, check the troubleshooting section below.

06_evaluation.py

Runs evaluation and testing patterns for LLM outputs.

python examples/06_evaluation.py

This script may use criteria-based evaluation or comparison metrics. Adjust the evaluation criteria inside the script to fit your use case.

07_memory.py

Shows conversation history management with memory.

python examples/07_memory.py

The script maintains a chat history and uses it to provide context in a multi-turn conversation.

08_batch_inference.py

Demonstrates batch inference for processing multiple inputs efficiently.

python examples/08_batch_inference.py

This script shows how to send multiple prompts in a single API call using batch endpoints, reducing latency and cost.

09_parallel_tool_calls.py

Demonstrates parallel tool calling, where the model requests multiple tool invocations in a single response.

python examples/09_parallel_tool_calls.py

This example shows how to handle multiple tool calls from one model response, execute them concurrently, and feed the results back to the model.

10_async.py

Demonstrates async/await patterns for concurrent model calls.

python examples/10_async.py

This example shows how to run multiple model calls concurrently using asyncio and LangChain's async methods.

11_parallel_tool_calls.py

Demonstrates parallel tool calls with async execution.

python examples/11_parallel_tool_calls.py

This example shows how to combine async/await with parallel tool calls for concurrent tool execution.

Running Examples

# Run any example
python examples/01_basic_chains.py

# Or with specific model (overrides .env)
LANGCHAIN_MODEL=anthropic/claude-3-5-sonnet python examples/01_basic_chains.py

Troubleshooting

Issue	Solution
ModuleNotFoundError: langchain_anthropic	pip install langchain-anthropic
ModuleNotFoundError: langchain_openai	pip install langchain-openai
ModuleNotFoundError: langchain_google_genai	pip install langchain-google-genai
ModuleNotFoundError: langchain_groq	pip install langchain-groq
ModuleNotFoundError: langchain_ollama	pip install langchain-ollama
AuthenticationError (Anthropic)	Verify ANTHROPIC_API_KEY in .env; key must start with sk-ant-; check console.anthropic.com for valid key
AuthenticationError (OpenAI)	Verify OPENAI_API_KEY in .env; key must start with sk-; check platform.openai.com for valid key and billing
AuthenticationError (Google)	Verify GOOGLE_API_KEY in .env; enable Generative Language API in Google Cloud Console; key must have API access
AuthenticationError (Groq)	Verify GROQ_API_KEY in .env; key must start with gsk_; check console.groq.com for valid key
init_chat_model returns wrong model	Ensure LANGCHAIN_MODEL uses format provider/model-name (e.g., anthropic/claude-3-5-sonnet, openai/gpt-4o, google/gemini-1.5-pro, groq/llama-3.1-70b-versatile, ollama/llama3.1)
Streaming not working	Confirm model supports streaming; some providers require streaming=True in model kwargs; Ollama requires streaming=True explicitly
Ollama connection refused	Start Ollama server: ollama serve and pull model: `ollama