流式訊息

在建立訊息時，您可以設置 "stream": true 來使用 server-sent events (SSE) 逐步流式傳輸回應。

使用 SDK 進行流式傳輸

我們的 Python 和 Typescript SDK 提供多種流式傳輸方式。Python SDK 允許同步和異步流。詳情請參閱每個 SDK 的文檔。

import anthropic

client = anthropic.Anthropic()

with client.messages.stream(
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
    model="claude-3-opus-20240229",
) as stream:
  for text in stream.text_stream:
      print(text, end="", flush=True)

事件類型

每個 server-sent 事件都包含一個命名的事件類型和相關的 JSON 數據。每個事件都將使用 SSE 事件名稱（例如 event: message_stop），並在其數據中包含匹配的事件 type。

每個流使用以下事件流程：

message_start：包含一個 Message 對象，其中 content 為空。
一系列內容塊，每個內容塊都有一個 content_block_start、一個或多個 content_block_delta 事件和一個 content_block_stop 事件。每個內容塊都有一個 index，對應於其在最終 Message content 數組中的索引。
一個或多個 message_delta 事件，表示對最終 Message 對象的頂層更改。
最後一個 message_stop 事件。

Ping 事件

事件流中還可能包含任意數量的 ping 事件。

錯誤事件

我們偶爾可能會在事件流中發送錯誤。例如，在使用量高的時期，您可能會收到一個 overloaded_error，這通常對應於非流式上下文中的 HTTP 529：

錯誤示例

event: error
data: {"type": "error", "error": {"type": "overloaded_error", "message": "Overloaded"}}

其他事件

根據我們的版本控制策略，我們可能會添加新的事件類型，您的代碼應該優雅地處理未知的事件類型。

Delta 類型

每個 content_block_delta 事件都包含一個 delta，其類型更新給定 index 處的 content 塊。

文本 delta

text 內容塊 delta 如下所示：

文本 delta

event: content_block_delta
data: {"type": "content_block_delta","index": 0,"delta": {"type": "text_delta", "text": "ello frien"}}

輸入 JSON delta

tool_use 內容塊的 delta 對應於塊的 input 字段的更新。為了支持最大的粒度，delta 是 部分 JSON 字符串，而最終的 tool_use.input 始終是一個對象。

您可以累積字符串 delta，並在收到 content_block_stop 事件時使用像 Pydantic 這樣的庫進行部分 JSON 解析，或者使用我們的 SDK，它們提供了訪問解析後的增量值的幫助程序。

tool_use 內容塊 delta 如下所示：

輸入 JSON delta

event: content_block_delta
data: {"type": "content_block_delta","index": 1,"delta": {"type": "input_json_delta","partial_json": "{\"location\": \"San Fra"}}}

注意：我們當前的模型一次只支持從 input 中發出一個完整的鍵和值屬性。因此，在使用工具時，在流式事件之間可能會有延遲，而模型正在工作。一旦累積了一個 input 鍵和值，我們就會將它們作為多個 content_block_delta 事件發出，並使用分塊的部分 json，以便該格式可以自動支持未來模型中更細粒度的控制。

原始 HTTP 流響應

我們強烈建議在使用流式模式時使用我們的客戶端 SDK。但是，如果您正在構建直接的 API 集成，則需要自己處理這些事件。

流響應由以下部分組成：

一個 message_start 事件
可能有多個內容塊，每個內容塊包含： a. 一個 content_block_start 事件 b. 可能有多個 content_block_delta 事件 c. 一個 content_block_stop 事件
一個 message_delta 事件
一個 message_stop 事件

響應中可能還會分散 ping 事件。有關格式的更多詳細信息，請參閱事件類型。

基本流請求

請求

curl https://api.anthropic.com/v1/messages \
     --header "anthropic-version: 2023-06-01" \
     --header "content-type: application/json" \
     --header "x-api-key: $ANTHROPIC_API_KEY" \
     --data \
'{
  "model": "claude-3-opus-20240229",
  "messages": [{"role": "user", "content": "Hello"}],
  "max_tokens": 256,
  "stream": true
}'

響應

event: message_start
data: {"type": "message_start", "message": {"id": "msg_1nZdL29xx5MUA1yADyHTEsnR8uuvGzszyY", "type": "message", "role": "assistant", "content": [], "model": "claude-3-opus-20240229", "stop_reason": null, "stop_sequence": null, "usage": {"input_tokens": 25, "output_tokens": 1}}}

event: content_block_start
data: {"type": "content_block_start", "index": 0, "content_block": {"type": "text", "text": ""}}

event: ping
data: {"type": "ping"}

event: content_block_delta
data: {"type": "content_block_delta", "index": 0, "delta": {"type": "text_delta", "text": "Hello"}}

event: content_block_delta
data: {"type": "content_block_delta", "index": 0, "delta": {"type": "text_delta", "text": "!"}}

event: content_block_stop
data: {"type": "content_block_stop", "index": 0}

event: message_delta
data: {"type": "message_delta", "delta": {"stop_reason": "end_turn", "stop_sequence":null}, "usage": {"output_tokens": 15}}

event: message_stop
data: {"type": "message_stop"}

使用工具的流請求

在此請求中，我們要求 Claude 使用工具告訴我們天氣。

請求

  curl https://api.anthropic.com/v1/messages \
    -H "content-type: application/json" \
    -H "x-api-key: $ANTHROPIC_API_KEY" \
    -H "anthropic-version: 2023-06-01" \
    -d '{
      "model": "claude-3-opus-20240229",
      "max_tokens": 1024,
      "tools": [
        {
          "name": "get_weather",
          "description": "Get the current weather in a given location",
          "input_schema": {
            "type": "object",
            "properties": {
              "location": {
                "type": "string",
                "description": "The city and state, e.g. San Francisco, CA"
              }
            },
            "required": ["location"]
          }
        }
      ],
      "tool_choice": {"type": "any"},
      "messages": [
        {
          "role": "user",
          "content": "What is the weather like in San Francisco?"
        }
      ],
      "stream": true
    }'

響應

event: message_start
data: {"type":"message_start","message":{"id":"msg_014p7gG3wDgGV9EUtLvnow3U","type":"message","role":"assistant","model":"claude-3-haiku-20240307","stop_sequence":null,"usage":{"input_tokens":472,"output_tokens":2},"content":[],"stop_reason":null}}

event: content_block_start
data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}

event: ping
data: {"type": "ping"}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"Okay"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":","}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" let"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":"'s"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" check"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" the"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" weather"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" for"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" San"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" Francisco"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":","}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":" CA"}}

event: content_block_delta
data: {"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":":"}}

event: content_block_stop
data: {"type":"content_block_stop","index":0}

event: content_block_start
data: {"type":"content_block_start","index":1,"content_block":{"type":"tool_use","id":"toolu_01T1x1fJ34qAmk2tNTrN7Up6","name":"get_weather","input":{}}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":""}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"{\"location\":"}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":" \"San"}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":" Francisc"}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"o,"}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":" CA\""}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":", "}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"\"unit\": \"fah"}}

event: content_block_delta
data: {"type":"content_block_delta","index":1,"delta":{"type":"input_json_delta","partial_json":"renheit\"}"}}

event: content_block_stop
data: {"type":"content_block_stop","index":1}

event: message_delta
data: {"type":"message_delta","delta":{"stop_reason":"tool_use","stop_sequence":null},"usage":{"output_tokens":89}}

event: message_stop
data: {"type":"message_stop"}

使用 API

Anthropic APIs

Amazon Bedrock API

Vertex AI

使用 SDK 進行流式傳輸

事件類型

Ping 事件

錯誤事件

其他事件

Delta 類型

文本 delta

輸入 JSON delta

原始 HTTP 流響應

基本流請求

使用工具的流請求

使用 API

Anthropic APIs

Amazon Bedrock API

Vertex AI

​使用 SDK 進行流式傳輸

​事件類型

​Ping 事件

​錯誤事件

​其他事件

​Delta 類型

​文本 delta

​輸入 JSON delta

​原始 HTTP 流響應

​基本流請求

​使用工具的流請求

使用 SDK 進行流式傳輸

事件類型

Ping 事件

錯誤事件

其他事件

Delta 類型

文本 delta

輸入 JSON delta

原始 HTTP 流響應

基本流請求

使用工具的流請求