Your First API Call

Adding AI to an app means calling a model provider's API from your server: you send messages, it returns generated content. This track uses Anthropic's Claude API with the official Python SDK, inside a Flask app. The same architecture (server-side calls, secrets in env vars, streaming, tools, structured output, cost control) applies to OpenAI and other providers through their own SDKs.

Get a key and keep it secret

Create an API key in the provider console and store it as an environment variable. Never put it in frontend JavaScript, Git or client apps: anyone could read it and spend your money.

export ANTHROPIC_API_KEY="sk-ant-..."     # locally; on Render/Railway use the dashboard
pip install anthropic

A first request

import anthropic

client = anthropic.Anthropic()   # reads ANTHROPIC_API_KEY from the environment

response = client.messages.create(
    model="claude-opus-5-5",
    max_tokens=16000,
    system="You are a concise assistant for a web development learning platform.",
    messages=[{"role": "user", "content": "Explain what a REST API is in two sentences."}],
)

for block in response.content:
    if block.type == "text":
        print(block.text)

print(response.usage.input_tokens, response.usage.output_tokens)

• model: which model to use. Model IDs change as new versions ship; keep the ID in configuration, not scattered through code.
• system: stable instructions for the whole conversation.
• messages: the conversation, alternating user and assistant.
• response.content is a list of blocks (text, thinking, tool use). Check each block's type before reading .text.
• usage tells you tokens used, which is what you pay for.

Choosing a model

Prices and model lists change; check the provider's model and pricing pages when you build.

Conversations are stateless

The API does not remember previous calls. To continue a chat, send the full history each time:

history = [
    {"role": "user", "content": "My name is Ada."},
    {"role": "assistant", "content": "Nice to meet you, Ada!"},
    {"role": "user", "content": "What is my name?"},
]
client.messages.create(model="claude-opus-5-5", max_tokens=16000, messages=history)

Store history in your database per conversation and trim or summarise it when it grows long.

Exercise

Write a script that asks the user a programming question, sends it with a helpful system prompt, prints the answer and the token usage, and keeps the conversation going for three turns.

Check yourself

• Why must the API key stay on the server?
• Why check block.type before reading text?
• How does the model "remember" earlier messages?