The Claude API is the door to automation: automatic reports, email processing, moderation, content generation at scale. This guide takes you from zero to your first working automation.

Advertisement

Getting started

You need a console.anthropic.com account and an API key. Install the official SDK:

# Python
pip install anthropic
# Node.js
npm install @anthropic-ai/sdk

Your first call

from anthropic import Anthropic
client = Anthropic(api_key="your-key")
message = client.messages.create(
    model="claude-sonnet-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Summarize in 3 bullets: [text]"}]
)
print(message.content[0].text)

Pattern 1: batch processing

import csv
from anthropic import Anthropic
client = Anthropic()
with open("emails.csv") as f:
    for row in csv.DictReader(f):
        response = client.messages.create(
            model="claude-haiku-4-5-20251001",
            max_tokens=200,
            messages=[{"role": "user", "content": f"Classify: {row['body']}"}]
        )
        print(row['id'], response.content[0].text)

Pattern 2: agent with tool use

tools = [{
    "name": "get_weather",
    "description": "Returns current weather for a city",
    "input_schema": {
        "type": "object",
        "properties": {"city": {"type": "string"}},
        "required": ["city"]
    }
}]
response = client.messages.create(
    model="claude-sonnet-5",
    max_tokens=1024,
    tools=tools,
    messages=[{"role": "user", "content": "What's the weather in Miami?"}]
)

Pattern 3: streaming for fast UX

with client.messages.stream(
    model="claude-sonnet-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "..."}]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

Pattern 4: prompt caching

messages = [{
    "role": "user",
    "content": [
        {"type": "text", "text": long_manual, "cache_control": {"type": "ephemeral"}},
        {"type": "text", "text": user_question}
    ]
}]
Advertisement

Pattern 5: webhooks

Receive external events (new email, form, ticket) and respond with Claude automatically. Combine API with Zapier/Make, AWS Lambda, or n8n.

Typical mistakes to avoid

  • API key on frontend: NEVER. Only from backend.
  • No retry or backoff.
  • Too low max_tokens.
  • No logging.

Cost and monitoring

Set budgets in Anthropic console. Log tokens per request. Use Haiku for tasks not needing Sonnet.

Conclusion

Claude API is the door to real automation. Start small: a script processing 50 items automatically gives you an idea of the power. Scale from there with confidence.

Keep exploring