Claude API

Build applications with Anthropic's Claude API using official Python and TypeScript SDKs.

Base URL: https://api.anthropic.com

Authentication

All requests require these headers (SDKs handle automatically):

Quick Start

Python

from anthropic import Anthropic

client = Anthropic()  # Uses ANTHROPIC_API_KEY env var

message = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello, Claude"}]
)
print(message.content[0].text)

TypeScript

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic();  // Uses ANTHROPIC_API_KEY env var

const message = await client.messages.create({
  model: "claude-sonnet-4-5-20250929",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello, Claude" }],
});
console.log(message.content[0].text);

Available Models

For complete model details, platform IDs (AWS Bedrock, GCP Vertex AI), and legacy models, see [references/models.md](references/models.md).

Available APIs

For full API details, rate limits, and third-party platforms, see [references/api-overview.md](references/api-overview.md).

Core Capabilities

Messages API

Basic chat completions, multi-turn conversations, system prompts, streaming responses.

Tool Use

Define tools with JSON schemas, handle tool calls, return results to Claude.

Vision

Send images (base64 or URL), analyze multiple images, extract information.

Streaming

Real-time token streaming with event handlers or async iterators.

Token Counting

Count tokens before sending requests to manage costs and rate limits.

API & SDK References

• API Overview: See [references/api-overview.md](references/api-overview.md) - endpoints, authentication, rate limits, third-party platforms
• Messages API: See [references/messages-api.md](references/messages-api.md) - request/response structure, content types
• Vision: See [references/vision.md](references/vision.md) - image inputs, base64/URL/Files API sources, limits
• Streaming: See [references/streaming.md](references/streaming.md) - SSE events, deltas, error recovery
• Python SDK: See [references/python-sdk.md](references/python-sdk.md)
• TypeScript SDK: See [references/typescript-sdk.md](references/typescript-sdk.md)

Advanced Features

• Extended Thinking: See [references/extended-thinking.md](references/extended-thinking.md) - step-by-step reasoning, tool use, interleaved thinking
• Citations: See [references/citations.md](references/citations.md) - document citations, source verification
• Token Counting: See [references/token-counting.md](references/token-counting.md) - count tokens before sending, free endpoint
• Context Windows: See [references/context-windows.md](references/context-windows.md) - 200K/1M context, context awareness
• Prompt Caching: See [references/prompt-caching.md](references/prompt-caching.md) - reduce latency and costs
• Batches, PDFs: See [references/advanced-features.md](references/advanced-features.md)

Common Patterns

Conversation Loop with Tools

messages = [{"role": "user", "content": user_input}]

while True:
    response = client.messages.create(
        model="claude-sonnet-4-5-20250929",
        max_tokens=1024,
        tools=tools,
        messages=messages
    )

    if response.stop_reason == "end_turn":
        break

    if response.stop_reason == "tool_use":
        messages.append({"role": "assistant", "content": response.content})
        tool_results = []
        for block in response.content:
            if block.type == "tool_use":
                result = execute_tool(block.name, block.input)
                tool_results.append({
                    "type": "tool_result",
                    "tool_use_id": block.id,
                    "content": result
                })
        messages.append({"role": "user", "content": tool_results})

Error Handling with Retries

from anthropic import RateLimitError
import time

def call_with_retry(fn, max_retries=3):
    for attempt in range(max_retries):
        try:
            return fn()
        except RateLimitError as e:
            if attempt == max_retries - 1:
                raise
            wait = int(e.response.headers.get("retry-after", 60))
            time.sleep(wait)

Required Parameters

Every messages.create() call requires: