from langchain.prompts import PromptTemplate` raises ImportError in langchain>=1.x

[TecheraFabrizio]

Description

Importing PromptTemplate from langchain.prompts fails with a module resolution error in recent versions of LangChain.

Steps to reproduce

1. Install langchain 1.2.10
2. Run the following code:

from langchain.prompts import PromptTemplate

Expected behavior

PromptTemplate is importable from langchain.prompts as documented in most existing tutorials and examples.

Actual behavior

Module not found / cannot be resolved error.

Workaround

Use the following import instead:

from langchain_core.prompts import PromptTemplate

Environment

• langchain: 1.2.10
• langchain-core: 1.2.15
• Python: 3.13
• OS: Windows 11

Notes

Most tutorials and official examples still reference langchain.prompts, which leads to confusion for new users.
Either the old import path should be maintained for backwards compatibility, or the documentation and examples should be updated to reflect the new structure.

[xXMrNidaXx]

This is a common issue with LangChain's package restructuring. Here's a complete migration guide:

Old vs New Imports:

# ❌ Old (deprecated)
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain
from langchain.llms import OpenAI

# ✅ New (langchain >= 0.2)
from langchain_core.prompts import PromptTemplate, ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser
from langchain_openai import ChatOpenAI  # separate package

Compatibility Shim (if you need to support both):

try:
    from langchain_core.prompts import PromptTemplate
except ImportError:
    from langchain.prompts import PromptTemplate

Key Changes:

• Core abstractions moved to langchain-core
• LLM providers are now separate packages (langchain-openai, langchain-anthropic, etc.)
• LLMChain deprecated in favor of LCEL (LangChain Expression Language)

Modern LCEL Pattern:

from langchain_core.prompts import ChatPromptTemplate
from langchain_openai import ChatOpenAI
from langchain_core.output_parsers import StrOutputParser

prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a helpful assistant."),
    ("human", "{input}")
])

chain = prompt | ChatOpenAI(model="gpt-4o") | StrOutputParser()
result = chain.invoke({"input": "Hello!"})

The new structure is cleaner once you get used to it. The main benefit is smaller installs - you only need to install the LLM packages you actually use.

[TecheraFabrizio]

Yes indeed thanks!