langchain-python-quickstart

langchain-ai/langchain-skills

Scaffold a minimal LangChain Python agent locally from the official quickstart.

What is langchain-python-quickstart?

Quickly set up a working LangChain agent in Python by following the official quickstart docs. Use this when you want to build or test a basic agent locally without complex setup. The skill handles model selection, environment configuration, and runs a working example.

Fetches and implements the official LangChain Python quickstart
Prompts you to choose a model provider (OpenAI, Anthropic, Google, etc.)
Creates an isolated project directory to keep your workspace clean
Sets up environment variables for API keys in a .env file
Installs required provider packages automatically
Runs a working weather agent example and displays output

How to install langchain-python-quickstart

npx skills add https://github.com/langchain-ai/langchain-skills --skill langchain-python-quickstart

Prerequisites

Python 3.8 or later installed
An API key for your chosen model provider (OpenAI, Anthropic, Google GenAI, etc.)
Basic familiarity with command-line tools

Claude Code

Cursor

Windsurf

Cline

How to use langchain-python-quickstart

1.Run the skill and choose your model provider when prompted (e.g., anthropic:claude-sonnet-5)
2.Provide your API key in the generated .env file
3.The skill creates a new langchain-agent/ directory with all scaffolding
4.Review the generated agent code and run it to see the weather agent in action
5.Check the output to verify the agent is working, then explore langchain-fundamentals for next steps

Use cases

Good for

Getting started with LangChain for the first time
Testing LangChain with different model providers before committing to a larger project
Creating a minimal agent scaffold to build upon
Learning how LangChain's create_agent function works
Quickly prototyping agent behavior locally

Who it's for

Python developers new to LangChain
Engineers evaluating LangChain for a project
Anyone wanting a quick, model-agnostic agent setup
Developers learning agentic AI patterns

langchain-python-quickstart FAQ

Do I need LangSmith or Tavily API keys?

No. This quickstart works without them. You only need an API key for your chosen model provider.

Can I use a different model than the default?

Yes. The skill prompts you to choose any provider:model combination (e.g., openai:gpt-4, google_genai:gemini-2.5-flash-lite). It installs the required provider package automatically.

Where does my API key go?

It goes in a .env file in the project directory, which is gitignored. You should edit it yourself—the skill won't paste keys into chat.

What happens after the quickstart runs?

The skill shows you the output and points you to langchain-fundamentals for learning more advanced LangChain patterns.

Will this affect my existing projects?

No. The skill creates a new isolated langchain-agent/ directory, so your existing code stays untouched.

Full instructions (SKILL.md)

Source of truth, from langchain-ai/langchain-skills.

name: langchain-python-quickstart description: "Scaffold a minimal local LangChain agent in Python by following the official quickstart. Use when the user wants to quickly build or try a LangChain agent locally."

LangChain Python quickstart

Follow the live docs — do not invent an alternate API from memory:

https://docs.langchain.com/oss/python/langchain/quickstart

Fetch that page (Docs MCP or HTTP) and implement what it shows (weather agent + create_agent).

Local setup constraints

Apply these on top of the quickstart (they keep setup minimal and model-agnostic):

Ask which provider/model to use. Showcase that LangChain is model-agnostic. Suggested prompt:

Which model should this agent use? Pass a provider:model string — e.g. openai:gpt-5.5, anthropic:claude-sonnet-5, google_genai:gemini-2.5-flash-lite. Default if you're unsure: anthropic:claude-sonnet-5.

Swap the quickstart's model string for their choice (or the default).
Create a new directory (e.g. langchain-agent/) and do all work there — do not pollute the open project.
Only secret: the provider API key in .env (gitignored). No LangSmith / Tavily unless they ask. Prefer they edit .env themselves — don't paste keys into chat.
Install the provider package needed for their model if the quickstart's base install isn't enough.
Run the example, show output, then stop. Point to langchain-fundamentals for next steps.