Using Claude Projects and Extended Thinking

mkdir my-claude-project
cd my-claude-project
claude init

claude upload docs/project-setup.md

{
  "model": "claude-3-5-sonnet",
  "messages": [
    {
      "role": "user",
      "content": "What is the capital of France?"
    }
  ],
  "tools": [
    {
      "name": "extended_thinking",
      "type": "claude_extended_thinking"
    }
  ]
}

const model = process.env.MODEL || "claude-3-opus"; // or sonnet, haiku

if (model === "claude-3-opus") {
  // Use for complex reasoning and code
} else if (model === "claude-3-sonnet") {
  // Use for balanced performance
} else if (model === "claude-3-haiku") {
  // Use for fast responses
}

claude artifacts create --name "my-artifact" --description "A sample artifact for code preview"

const claudeClient = new ClaudeClient({
  maxContextLength: 100,
  systemPrompt: "You are a senior software engineer specializing in React and Node.js development.",
  temperature: 0.7
});

// Example usage with conversation history
const response = await claudeClient.sendMessage({
  messages: [
    { role: "user", content: "How do I implement a login flow in React?" },
    { role: "assistant", content: "To implement a login flow in React..." }
  ],
  maxTokens: 500
});

claude --computer-use --file-op --web-browse

const Claude = require('claude-sdk');

const client = new Claude({
  system_prompt: "You are an expert JavaScript developer who explains concepts clearly.",
  max_tokens: 500,
  temperature: 0.5
});

client.sendMessage("Explain closures in JavaScript");

// Estimate tokens in a string using Claude's tokenizer
const inputText = "This is a sample text to estimate token count.";
// Using Claude's tokenizer (e.g., via the client library)
const tokenCount = tokenize(inputText).length;
console.log(`Token count: ${tokenCount}`);