LangChain Python

May 25, 2023 1 min read python, llm, genai

Lab note What I wrote down while learning this by building with it. Short by design; the code is in pfeilbr/langchain-python-playground.

learn LangChain (python)

concepts

Agents - use an LLM to determine which actions to take and in what order. An action can either be using a tool and observing its output, or returning to the user.

the llm is used to select the tool to use

langchain/agents/agent.py - _call method has main logic langchain/agents/types.py langchain/agents/mrkl/prompt.py langchain/agents/**/*prompt.py - these contain all the prompts for different agents

demo

pipenv shell
export OPENAI_API_KEY=...
export SERPAPI_API_KEY=...
jupyter lab
# visit `main.ipynb`

resources

langchain-python
LangChain AI Handbook - great resource

Keep learning

Free learning path · 19 resources Get good with a specific AI model Each lab's own courses, cookbooks and docs — Claude, ChatGPT, Gemini, Llama, Mistral — and how to run open models yourself.
Course · 12 lessons mlx-lm Run, quantize, serve, and fine-tune LLMs locally on Apple Silicon with Apple's mlx-lm — the CLI and Python surfaces, from first principles, …
Course · 11 lessons Agentic Loop Engineering Designing tight, verifiable agent loops with Claude Code — the verify step, tools, and feedback.

concepts

demo

resources

Keep learning

Get new guides and courses by email