Tutorial on How to Use Claude: Step-by-Step Guide

If you want a tutorial on how to use Claude, there are really two paths depending on what you're trying to do. If you want a chat assistant for writing, research, or coding help, you use Claude through the web app or desktop/mobile apps at claude.ai. If you want to build Claude into a product, script, or automation, you use it through an API and send requests programmatically. This guide covers both, plus the practical details that trip people up: setting up conversations correctly, structuring API calls, and getting Claude to do what you actually want.

By the end you'll know how to start a productive Claude conversation in the browser, how to make your first API call, and how to decide which approach fits your use case.

Using Claude in the Web App

The fastest way to use Claude is at claude.ai. Sign up, and you land in a chat interface that works like any modern chatbot, with a few things worth knowing:

• Projects let you group related conversations and attach reference documents (style guides, codebases, specs) that Claude reads before every reply in that project.
• Artifacts open a side panel for code, documents, or diagrams Claude generates, so you can iterate on a single piece of content without it getting buried in chat history.
• File uploads accept PDFs, images, spreadsheets, and code files directly in the conversation.

A basic session looks like this: open a new chat, give Claude context (what you're building, who the audience is, any constraints), then ask your actual question. Claude performs noticeably better when you front-load context instead of trickling it in one message at a time.

Prompting tips that actually move the needle

• Be specific about format. "Write three bullet points, each under 15 words" gets a cleaner result than "summarize this."
• Give examples. If you want a specific tone or structure, paste one example and say "match this style."
• Ask for reasoning when it matters. For anything involving math, logic, or multi-step decisions, ask Claude to think step by step before giving a final answer — it catches its own mistakes more often this way.
• Iterate instead of restarting. If the first answer is close but not right, tell Claude what's wrong specifically rather than rephrasing the whole prompt from scratch.

Using Claude Through the API

Once you need Claude inside an application — a support bot, a content pipeline, an internal tool — the chat interface isn't the right layer. You need the API, which takes a structured request and returns a structured response.

A minimal request looks like this:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Explain what a race condition is in one paragraph."}
    ]
  }'

The response comes back as JSON with the generated text, token usage, and a stop reason. From there you build conversations by appending each turn to the messages array, with alternating user and assistant roles. That's the entire mental model: you're not managing a session on Anthropic's servers, you're sending the full conversation history with every request.

For anything beyond a single question, you'll also want to control max_tokens (cost and response length), temperature (higher for creative tasks, lower for factual or code-related ones), and system prompts, which set persistent instructions that apply to the whole conversation without cluttering the message history.

Getting API Access Without a Separate Anthropic Account

Here's a detail a lot of tutorials skip: to call the Anthropic API directly, you typically need a separate developer account with its own billing, distinct from a personal or team Claude.ai subscription. That's fine if you're building a standalone product, but it's redundant if you already pay for Claude and just want programmatic access for a script, an internal tool, or a small app.

SubToAPI exists for exactly this gap. It turns your existing Claude access into a standard HTTPS API: you get application-scoped keys (sub_live_...), streaming responses, tool use, usage metadata per key, and team seat management, all from one dashboard, without setting up separate API billing.

A request against SubToAPI looks almost identical to a direct API call:

const response = await fetch("https://api.subtoapi.app/v1/messages", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.SUBTOAPI_KEY}`,
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    model: "claude-sonnet-4-5",
    max_tokens: 1024,
    messages: [
      { role: "user", content: "Write a regex that matches US phone numbers." }
    ]
  })
});

const data = await response.json();
console.log(data);

If you're just testing things out, start with the quickstart, which walks through creating a key and making your first call. The messages docs cover the full request/response shape, streaming covers token-by-token output for chat UIs, and tools covers function calling if you need Claude to invoke your own APIs mid-conversation. Plans start at Solo for €9/month, with Team (€19/seat) and Scale (€49/seat) tiers for multi-key or multi-user setups — see pricing for details, and there's a free trial at signup.

Choosing Between Web App and API

Use the web app when you're doing one-off work: writing, debugging a snippet, researching a topic, or reviewing a document. Use the API — direct or through SubToAPI — when the interaction needs to happen without a human clicking "send" each time: chatbots, batch document processing, automated code review, content generation pipelines. If you're not sure yet, start in the web app to validate that Claude produces the output quality you need, then move the same prompts into API calls once you know the shape of the task.

questions

Do I need to know how to code to use Claude? No. The web app at claude.ai works like any chat interface — no coding required. Coding is only necessary if you're integrating Claude into an application via the API.

What's the difference between using Claude in the browser and through the API? The browser gives you a persistent conversation UI, file uploads, and artifacts for free. The API gives you programmatic control: you send the full message history with each request and get structured JSON back, which you then build into your own app or workflow.

Can I use my existing Claude subscription to make API calls? Not directly through Anthropic, since API access normally requires separate developer billing. Services like SubToAPI bridge this by turning your Claude access into an API with its own keys, so you don't need to set up a second account — see the quickstart to get started.