概述
Anthropic Claude SDK for Python提供从Python应用程序访问Claude API的能力。
安装
pip install anthropic快速开始
import os
from anthropic import Anthropic
client = Anthropic(
api_key=os.environ.get("ANTHROPIC_API_KEY")
)
message = client.messages.create(
max_tokens=1024,
messages=[
{
"role": "user",
"content": "Hello, Claude"
}
],
model="claude-opus-4-6",
)
print(message.content)重要参数
模型选择
| 模型 | 描述 | 适合场景 |
|---|---|---|
| claude-opus-4-6 | 最强性能 | 复杂推理 |
| claude-sonnet-4-6 | 平衡 | 日常任务 |
| claude-haiku-3-5 | 最快 | 简单任务 |
max_tokens
生成的最大token数。设置较大值可获得更长输出：
message = client.messages.create(
max_tokens=4096,
messages=[...],
model="claude-opus-4-6"
)系统提示词
message = client.messages.create(
max_tokens=1024,
system="你是一个专业的AI助手，擅长技术写作",
messages=[
{"role": "user", "content": "解释什么是机器学习"}
],
model="claude-opus-4-6"
)图像输入（Claude 3.5+）
message = client.messages.create(
max_tokens=1024,
messages=[
{
"role": "user",
"content": [
{"type": "image", "source": {"type": "base64", "media_type": "image/jpeg", "data": "..."}},
{"type": "text", "text": "这张图片里有什么?"}
]
}
],
model="claude-opus-4-6"
)流式响应
with client.messages.stream(
max_tokens=1024,
messages=[{"role": "user", "content": "写一个故事"}],
model="claude-sonnet-4-6"
) as stream:
for text in stream.text_deltas:
print(text, end="")工具使用（Function Calling）
from anthropic import Anthropic
client = Anthropic()
tools = [
{
"name": "get_weather",
"description": "获取指定城市的天气",
"input_schema": {
"type": "object",
"properties": {
"city": {"type": "string", "description": "城市名称"}
},
"required": ["city"]
}
}
]
message = client.messages.create(
max_tokens=1024,
tools=tools,
messages=[
{"role": "user", "content": "北京天气怎么样?"}
],
model="claude-opus-4-6"
)
if message.stop_reason == "tool_use":
tool_call = message.tool_calls[0]
if tool_call.name == "get_weather":
result = get_weather(tool_call.input["city"])Thinking模式
Claude 4.0+支持思考过程：
message = client.messages.create(
max_tokens=4096,
thinking={"type": "enabled", "budget_tokens": 4096},
messages=[{"role": "user", "content": "计算123*456"}],
model="claude-opus-4-6"
)
print(message.thinking)
print(message.content)错误处理
from anthropic import Anthropic, RateLimitError, APIStatusError
client = Anthropic()
try:
message = client.messages.create(...)
except RateLimitError:
print("速率超限")
except APIStatusError as e:
print(f"API错误: {e.status_code} - {e.message}")Python版本要求
Python 3.9+