langchain-fundamentals

1<oneliner>
2Build production agents using `create_agent()`, middleware patterns, and the `@tool` decorator / `tool()` function. When creating LangChain agents, you MUST use create_agent(), with middleware for custom flows. All other alternatives are outdated.
3</oneliner>
4 
5<create_agent>
6## Creating Agents with create_agent
7 
8`create_agent()` is the recommended way to build agents. It handles the agent loop, tool execution, and state management.
9 
10### Agent Configuration Options
11 
12| Parameter | Purpose | Example |
13|-----------|---------|---------|
14| `model` | LLM to use | `"anthropic:claude-sonnet-4-5"` or model instance |
15| `tools` | List of tools | `[search, calculator]` |
16| `system_prompt` / `systemPrompt` | Agent instructions | `"You are a helpful assistant"` |
17| `checkpointer` | State persistence | `MemorySaver()` |
18| `middleware` | Processing hooks | `[HumanInTheLoopMiddleware]` (Python) / `[humanInTheLoopMiddleware({...})]` (TypeScript) |
19</create_agent>
20 
21<ex-basic-agent>
22<python>
23 
24```python
25from langchain.agents import create_agent
26from langchain_core.tools import tool
27 
28@tool
29def get_weather(location: str) -> str:
30    """Get current weather for a location.
31 
32    Args:
33        location: City name
34    """
35    return f"Weather in {location}: Sunny, 72F"
36 
37agent = create_agent(
38    model="anthropic:claude-sonnet-4-5",
39    tools=[get_weather],
40    system_prompt="You are a helpful assistant."
41)
42 
43result = agent.invoke({
44…