How to Use Claude: Chat, API, and Building Apps

Claude is Anthropic's family of AI models, and you can use it in three main ways: through the Claude.ai web and desktop apps for everyday chat and writing, through the Claude Developer API for building your own products, or through a proxy layer that turns a Claude subscription into a standard HTTPS API. Which one you need depends entirely on what you're trying to accomplish.

If you just want a smart assistant for writing, research, or brainstorming, the web app at claude.ai is the fastest way to get started — sign up, type a message, and go. If you're a developer who wants to embed Claude into a product, script, or internal tool, you need the API, and that's what the rest of this guide focuses on.

Using Claude Through the Web App

For most non-technical use cases, this is the right starting point:

1. Go to claude.ai and create an account.
2. Pick a model tier based on your plan (Free, Pro, or Max).
3. Start a conversation — Claude supports long context windows, so you can paste in documents, code files, or transcripts and ask questions about them.
4. Use Projects to keep related conversations and reference documents organized.
5. Attach files (PDFs, images, spreadsheets) directly in the chat for analysis.

This covers drafting emails, summarizing documents, debugging small code snippets, and general Q&A. It doesn't give you programmatic access — there's no way to call this from your own app or automate it at scale.

Using Claude Through the Developer API

Once you need Claude inside a product — a chatbot, a content pipeline, a coding assistant — you move to the API. Anthropic's Messages API is the core interface: you send a list of messages and get a completion back, optionally streamed token by token.

A basic 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-4-20250514",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Explain how binary search works"}
    ]
  }'

Key concepts you'll run into:

• System prompts set behavior and constraints before the conversation starts.
• Tool use lets Claude call functions you define — useful for fetching live data, querying a database, or executing code.
• Streaming returns partial output as it's generated, which matters for chat UIs where latency feels bad if the user stares at a blank screen.
• Usage metadata in each response tells you input/output token counts, which you need for cost tracking and rate limiting.

This gets you full control, but it also means managing API keys, billing, rate limits, and — separately — your Claude subscription if you're already paying for Pro or Max access personally.

Why Some Teams Use a Proxy Instead

A common friction point: you (or your company) already pay for a Claude plan, but the Developer API is billed separately, with its own usage-based pricing and its own key management. If you want to build an internal tool, a prototype, or a small product without setting up a second billing relationship, a service like SubToAPI bridges that gap.

SubToAPI turns your existing Claude access into a standard HTTPS API with application keys, so you can call it from code without juggling separate API billing:

const res = 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-4-20250514",
    max_tokens: 1024,
    messages: [{ role: "user", content: "Summarize this changelog in 3 bullets" }]
  })
});

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

It supports the same core building blocks developers expect — streaming, tool use, usage metadata — under one dashboard, with application keys formatted like sub_live_... so you can issue different keys per app or environment. Plans start at €9/month for solo use, with Team (€19/seat) and Scale (€49/seat) tiers for organizations that need multiple seats and shared usage visibility. There's a free trial at signup if you want to test it before committing.

Check the /docs/quickstart guide for a full walkthrough, /docs/messages for request formats, /docs/streaming for real-time output, and /docs/tools for function calling.

Choosing the Right Approach

• Just chatting or writing → use claude.ai directly.
• Building a product at scale with dedicated infrastructure → use Anthropic's Developer API directly.
• Prototyping, internal tools, or turning an existing subscription into programmatic access without separate API billing → a proxy like SubToAPI is often faster to set up.

Whichever path you pick, the underlying skill is the same: write clear, specific prompts, use system messages to set consistent behavior, and lean on tool use when Claude needs to act on real data rather than just generate text. Start small — a single script or endpoint — before wiring Claude into a production workflow.

Questions

Do I need to know how to code to use Claude? No. The claude.ai web and desktop apps require no coding — you type messages and get responses like any chat interface. Coding is only necessary if you want to integrate Claude into your own application or automate it.

What's the difference between Claude.ai and the Claude API? Claude.ai is a finished product for human conversations in a browser. The API is a programmatic interface for developers to send requests from their own code and get structured responses, including streaming and tool calls.

Can I use my existing Claude subscription to build an app without paying separately for API access? Not directly through Anthropic — the Developer API is billed separately from a Claude.ai subscription. Services like SubToAPI exist specifically to expose your Claude access as an HTTPS API with its own key management, avoiding a second billing setup. See /pricing for current plans.