Python for LLM Apps: From Beginner to Production

Python is the lingua franca of LLM development. This tutorial takes you from your first API call to production-grade code in 30 minutes.

Install

pip install openai anthropic

Your first API call

import openai

client = openai.OpenAI()  # uses OPENAI_API_KEY env var
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "What is Python?"}]
)
print(response.choices[0].message.content)

Streaming (for chatbots)

stream = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Write a haiku"}],
    stream=True
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Error handling

from openai import APIError, RateLimitError, APIConnectionError
import time

def safe_call(messages, max_retries=3):
    for attempt in range(max_retries):
        try:
            return client.chat.completions.create(
                model="gpt-4o-mini", messages=messages
            )
        except RateLimitError:
            time.sleep(2 ** attempt)  # exponential backoff
        except APIConnectionError as e:
            print(f"Connection error: {e}")
            time.sleep(5)
    raise Exception("All retries failed")

Async batch processing

import asyncio
from openai import AsyncOpenAI

client = AsyncOpenAI()

async def process(prompt):
    response = await client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}]
    )
    return response.choices[0].message.content

async def main():
    prompts = ["Q1", "Q2", "Q3", "Q4", "Q5"]
    results = await asyncio.gather(*[process(p) for p in prompts])
    return results

results = asyncio.run(main())
print(f"Got {len(results)} responses")

Key takeaways

• Always set OPENAI_API_KEY env var (never hardcode)
• Use streaming for user-facing UIs
• Always have retry + backoff for production