Lab Guide — Claude API Setup & First Calls

Load environment and create the client

from dotenv import load_dotenv
import os
import anthropic

load_dotenv()
client = anthropic.Anthropic()
# The SDK automatically reads ANTHROPIC_API_KEY from the environment

Make the API call

response = client.messages.create(
    model="claude-3-5-haiku-20241022",
    max_tokens=256,
    messages=[
        {"role": "user", "content": "Tell me a fun fact about large language models."}
    ]
)

Print the response text and usage

print("Response:")
print(response.content[0].text)
print()
print("Token usage:")
print(f"  Input tokens:  {response.usage.input_tokens}")
print(f"  Output tokens: {response.usage.output_tokens}")
print(f"  Stop reason:   {response.stop_reason}")

Implement the function

def stream_response(prompt: str) -> None:
    """Stream a response from Claude, printing each chunk as it arrives."""
    with client.messages.stream(
        model="claude-3-5-haiku-20241022",
        max_tokens=512,
        messages=[{"role": "user", "content": prompt}]
    ) as stream:
        for text in stream.text_stream:
            print(text, end="", flush=True)
    print()  # newline after stream ends

Test it with a longer prompt

stream_response("Explain quantum entanglement in simple terms.")

Compare streaming vs non-streaming latency

import time

prompt = "Write a haiku about APIs."

# Non-streaming: waits for full response
start = time.time()
response = client.messages.create(
    model="claude-3-5-haiku-20241022",
    max_tokens=64,
    messages=[{"role": "user", "content": prompt}]
)
non_stream_time = time.time() - start
print(f"Non-streaming: got full response in {non_stream_time:.2f}s")
print(response.content[0].text)
print()

# Streaming: first token arrives quickly
print("Streaming (notice first token arrives faster):")
start = time.time()
stream_response(prompt)
stream_total_time = time.time() - start
print(f"Streaming: total time {stream_total_time:.2f}s")

Implement the chat loop

def chat_loop() -> None:
    """Run an interactive multi-turn conversation with Claude."""
    history = []
    print("Chat with Claude (type 'quit' to exit)\n")

    while True:
        user_input = input("You: ").strip()
        if user_input.lower() == "quit":
            print("Goodbye!")
            break

        # Append user message to history
        history.append({"role": "user", "content": user_input})

        # Call the API with the full history
        response = client.messages.create(
            model="claude-3-5-haiku-20241022",
            max_tokens=512,
            messages=history
        )

        # Extract the reply and append it to history
        assistant_reply = response.content[0].text
        history.append({"role": "assistant", "content": assistant_reply})

        # Show the reply
        print(f"Claude: {assistant_reply}\n")

chat_loop()

Test that Claude remembers context

Try this conversation to verify memory works:

You: My name is Alex and I love Python.
Claude: Nice to meet you, Alex! Python is a fantastic language...

You: What is my name?
Claude: Your name is Alex, as you mentioned earlier!

You: What programming language do I love?
Claude: You mentioned that you love Python!

Optional: add a system prompt to give Claude a persona

response = client.messages.create(
    model="claude-3-5-haiku-20241022",
    max_tokens=512,
    system="You are a friendly Python tutor. Keep answers concise and practical.",
    messages=history
)