Claude 消息接口

• 完全兼容 Anthropic Claude Messages 原生协议（POST /v1/messages）
• 支持多轮对话、流式 SSE、工具调用、extended thinking
• 支持文本、图像等多模态内容
• 响应为上游原样透传，无 {code, data} 外层包装

请求示例

bash curl https://api.openveer.com/v1/messages \ -H "x-api-key: $API_KEY" \ -H "anthropic-version: 2025-10-01" \ -H "content-type: application/json" \ -d '{ "model": "claude-sonnet-4-6", "max_tokens": 1024, "messages": [ {"role": "user", "content": "你好，世界"} ] }'

```python
import anthropic

client = anthropic.Anthropic(
api_key="YOUR_API_KEY",
base_url="https://api.openveer.com"
)

message = client.messages.create(
model="claude-sonnet-4-6",
max_tokens=1024,
messages=[
{"role": "user", "content": "你好，世界"}
]
)

print(message.content)
```

```javascript
import Anthropic from '@anthropic-ai/sdk';

const client = new Anthropic({
apiKey: process.env.API_KEY,
baseURL: 'https://api.openveer.com'
});

const message = await client.messages.create({
model: 'claude-sonnet-4-6',
max_tokens: 1024,
messages: [
{ role: 'user', content: '你好，世界' }
]
});

console.log(message.content);
```

```go
package main

import (
"bytes"
"encoding/json"
"fmt"
"io/ioutil"
"net/http"
"os"
)

func main() {
url := "https://api.openveer.com/v1/messages"

  payload := map[string]interface{}{
      "model": "claude-sonnet-4-6",
      "max_tokens": 1024,
      "messages": []map[string]string{
          {
              "role":    "user",
              "content": "你好，世界",
          },
      },
  }

  jsonData, _ := json.Marshal(payload)

  req, _ := http.NewRequest("POST", url, bytes.NewBuffer(jsonData))
  req.Header.Set("x-api-key", os.Getenv("API_KEY"))
  req.Header.Set("anthropic-version", "2025-10-01")
  req.Header.Set("Content-Type", "application/json")

  client := &http.Client{}
  resp, err := client.Do(req)
  if err != nil {
      panic(err)
  }
  defer resp.Body.Close()

  body, _ := ioutil.ReadAll(resp.Body)
  fmt.Println(string(body))

}
```

```java
import java.net.http.HttpClient;
import java.net.http.HttpRequest;
import java.net.http.HttpResponse;
import java.net.URI;

public class Main {
public static void main(String[] args) throws Exception {
String url = "https://api.openveer.com/v1/messages";
String apiKey = System.getenv("API_KEY");

      String payload = """
      {
        "model": "claude-sonnet-4-6",
        "max_tokens": 1024,
        "messages": [
          {
            "role": "user",
            "content": "你好，世界"
          }
        ]
      }
      """;

      HttpClient client = HttpClient.newHttpClient();
      HttpRequest request = HttpRequest.newBuilder()
          .uri(URI.create(url))
          .header("x-api-key", apiKey)
          .header("anthropic-version", "2025-10-01")
          .header("Content-Type", "application/json")
          .POST(HttpRequest.BodyPublishers.ofString(payload))
          .build();

      HttpResponse<String> response = client.send(request,
          HttpResponse.BodyHandlers.ofString());

      System.out.println(response.body());
  }

}
```

```php
"claude-sonnet-4-6", "max_tokens" => 1024, "messages" => [ [ "role" => "user", "content" => "你好，世界" ] ] ]; $ch = curl_init($url); curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); curl_setopt($ch, CURLOPT_POST, true); curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($payload)); curl_setopt($ch, CURLOPT_HTTPHEADER, [ "x-api-key: " . $apiKey, "anthropic-version: 2025-10-01", "Content-Type: application/json" ]); $response = curl_exec($ch); curl_close($ch); echo $response; ?>

```

```ruby
require 'net/http'
require 'json'
require 'uri'

url = URI("https://api.openveer.com/v1/messages")
api_key = ENV['API_KEY']

payload = {
model: "claude-sonnet-4-6",
max_tokens: 1024,
messages: [
{
role: "user",
content: "你好，世界"
}
]
}

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["x-api-key"] = api_key
request["anthropic-version"] = "2025-10-01"
request["Content-Type"] = "application/json"
request.body = payload.to_json

response = http.request(request)
puts response.body
```

```swift
import Foundation

let url = URL(string: "https://api.openveer.com/v1/messages")!
let apiKey = ProcessInfo.processInfo.environment["API_KEY"] ?? ""

let payload: [String: Any] = [
"model": "claude-sonnet-4-6",
"max_tokens": 1024,
"messages": [
[
"role": "user",
"content": "你好，世界"
]
]
]

var request = URLRequest(url: url)
request.httpMethod = "POST"
request.setValue(apiKey, forHTTPHeaderField: "x-api-key")
request.setValue("2025-10-01", forHTTPHeaderField: "anthropic-version")
request.setValue("application/json", forHTTPHeaderField: "Content-Type")
request.httpBody = try? JSONSerialization.data(withJSONObject: payload)

let task = URLSession.shared.dataTask(with: request) { data, response, error in
if let error = error {
print("Error: (error)")
return
}

  if let data = data, let responseString = String(data: data, encoding: .utf8) {
      print(responseString)
  }

}

task.resume()
```

```csharp
using System;
using System.Net.Http;
using System.Text;
using System.Threading.Tasks;

class Program
{
static async Task Main(string[] args)
{
var url = "https://api.openveer.com/v1/messages";
var apiKey = Environment.GetEnvironmentVariable("API_KEY");

      var payload = @"{
          ""model"": ""claude-sonnet-4-6"",
          ""max_tokens"": 1024,
          ""messages"": [
              {
                  ""role"": ""user"",
                  ""content"": ""你好，世界""
              }
          ]
      }";

      using var client = new HttpClient();
      client.DefaultRequestHeaders.Add("x-api-key", apiKey);
      client.DefaultRequestHeaders.Add("anthropic-version", "2025-10-01");

      var content = new StringContent(payload, Encoding.UTF8, "application/json");
      var response = await client.PostAsync(url, content);
      var result = await response.Content.ReadAsStringAsync();

      Console.WriteLine(result);
  }

}
```

```c
#include
#include
#include

int main(void) {
CURL curl;
CURLcode res;
const char api_key = getenv("API_KEY");

  curl_global_init(CURL_GLOBAL_DEFAULT);
  curl = curl_easy_init();

  if(curl) {
      const char *url = "https://api.openveer.com/v1/messages";
      const char *payload = "{"
          "\"model\":\"claude-sonnet-4-6\","
          "\"max_tokens\":1024,"
          "\"messages\":[{\"role\":\"user\",\"content\":\"你好，世界\"}]"
      "}";

      char auth_header[256];
      snprintf(auth_header, sizeof(auth_header), "x-api-key: %s", api_key);

      struct curl_slist *headers = NULL;
      headers = curl_slist_append(headers, auth_header);
      headers = curl_slist_append(headers, "anthropic-version: 2025-10-01");
      headers = curl_slist_append(headers, "Content-Type: application/json");

      curl_easy_setopt(curl, CURLOPT_URL, url);
      curl_easy_setopt(curl, CURLOPT_POSTFIELDS, payload);
      curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);

      res = curl_easy_perform(curl);

      if(res != CURLE_OK) {
          fprintf(stderr, "curl_easy_perform() failed: %s\n",
                  curl_easy_strerror(res));
      }

      curl_slist_free_all(headers);
      curl_easy_cleanup(curl);
  }

  curl_global_cleanup();
  return 0;

}
```

```objectivec
#import

int main(int argc, const char * argv[]) {
@autoreleasepool {
NSURL url = [NSURL URLWithString:@"https://api.openveer.com/v1/messages"];
NSString apiKey = [NSProcessInfo processInfo].environment[@"API_KEY"];

      NSDictionary *payload = @{
          @"model": @"claude-sonnet-4-6",
          @"max_tokens": @1024,
          @"messages": @[
              @{
                  @"role": @"user",
                  @"content": @"你好，世界"
              }
          ]
      };

      NSError *error;
      NSData *jsonData = [NSJSONSerialization dataWithJSONObject:payload
                                                        options:0
                                                          error:&error];

      NSMutableURLRequest *request = [NSMutableURLRequest requestWithURL:url];
      [request setHTTPMethod:@"POST"];
      [request setValue:apiKey forHTTPHeaderField:@"x-api-key"];
      [request setValue:@"2025-10-01" forHTTPHeaderField:@"anthropic-version"];
      [request setValue:@"application/json" forHTTPHeaderField:@"Content-Type"];
      [request setHTTPBody:jsonData];

      NSURLSessionDataTask *task = [[NSURLSession sharedSession] 
          dataTaskWithRequest:request
          completionHandler:^(NSData *data, NSURLResponse *response, NSError *error) {
              if (error) {
                  NSLog(@"Error: %@", error);
                  return;
              }
              NSString *result = [[NSString alloc] initWithData:data 
                                                      encoding:NSUTF8StringEncoding];
              NSLog(@"%@", result);
          }];

      [task resume];
      [[NSRunLoop mainRunLoop] run];
  }
  return 0;

}
```

```ocaml
( Requires cohttp and yojson libraries )
open Lwt
open Cohttp
open Cohttp_lwt_unix

let url = "https://api.openveer.com/v1/messages"
let api_key = Sys.getenv "API_KEY"

let payload = {|{
"model": "claude-sonnet-4-6",
"max_tokens": 1024,
"messages": [
{
"role": "user",
"content": "你好，世界"
}
]
}|}

let () =
let headers = Header.init ()
|> fun h -> Header.add h "x-api-key" api_key
|> fun h -> Header.add h "anthropic-version" "2025-10-01"
|> fun h -> Header.add h "Content-Type" "application/json"
in
let body = Cohttp_lwt.Body.of_string payload in

let response = Client.post ~headers ~body (Uri.of_string url) >>= fun (resp, body) ->
  body |> Cohttp_lwt.Body.to_string >|= fun body_str ->
  print_endline body_str
in
Lwt_main.run response

```

```dart
import 'dart:convert';
import 'dart:io';
import 'package:http/http.dart' as http;

void main() async {
final url = Uri.parse('https://api.openveer.com/v1/messages');
final apiKey = Platform.environment['API_KEY'];

final payload = {
  'model': 'claude-sonnet-4-6',
  'max_tokens': 1024,
  'messages': [
    {
      'role': 'user',
      'content': '你好，世界'
    }
  ]
};

final response = await http.post(
  url,
  headers: {
    'x-api-key': apiKey!,
    'anthropic-version': '2025-10-01',
    'Content-Type': 'application/json',
  },
  body: jsonEncode(payload),
);

print(response.body);

}
```

```r
library(httr)
library(jsonlite)

url <- "https://api.openveer.com/v1/messages"
api_key <- Sys.getenv("API_KEY")

payload <- list(
model = "claude-sonnet-4-6",
max_tokens = 1024,
messages = list(
list(
role = "user",
content = "你好，世界"
)
)
)

response <- POST(
url,
add_headers(
x-api-key = api_key,
anthropic-version = "2025-10-01",
Content-Type = "application/json"
),
body = toJSON(payload, auto_unbox = TRUE),
encode = "raw"
)

cat(content(response, "text"))
```

响应示例

json { "model": "claude-sonnet-4-6", "id": "msg_011CdfeHuC728oxqaLrRNbcB", "type": "message", "role": "assistant", "content": [ { "type": "text", "text": "你好！我是Claude。很高兴见到你。" } ], "stop_reason": "end_turn", "stop_sequence": null, "stop_details": null, "usage": { "input_tokens": 12, "cache_creation_input_tokens": 0, "cache_read_input_tokens": 0, "cache_creation": { "ephemeral_5m_input_tokens": 0, "ephemeral_1h_input_tokens": 0 }, "output_tokens": 18, "service_tier": "standard", "inference_geo": "global" } }

json { "error": { "code": "model_not_found", "message": "model not found (request id: 20260803181903172761862QzohKPXF)", "param": "", "type": "apimart_error" } }

json { "error": { "code": "", "message": "无效的API密钥 (request id: 20260803181903172761862QzohKPXF)", "param": "", "type": "apimart_error" } }

json { "error": { "code": "", "message": "账户余额不足 (request id: 20260803181903172761862QzohKPXF)", "param": "", "type": "apimart_error" } }

json { "error": { "code": "", "message": "请求过于频繁 (request id: 20260803181903172761862QzohKPXF)", "param": "", "type": "apimart_error" } }

json { "error": { "code": "", "message": "服务器内部错误 (request id: 20260803181903172761862QzohKPXF)", "param": "", "type": "apimart_error" } }

Authorizations

鉴权支持两种方式，任选其一：

Body

单条用户消息示例：

json [{"role": "user", "content": "你好，Claude"}]

多轮对话示例：

json [ {"role": "user", "content": "你好"}, {"role": "assistant", "content": "你好！我是Claude。"}, {"role": "user", "content": "能解释一下AI吗？"} ]

预填充助手回复：

json [ {"role": "user", "content": "太阳的希腊名称是？(A) Sol (B) Helios (C) Sun"}, {"role": "assistant", "content": "答案是 ("} ]

Response

使用示例

基础对话

import anthropic

client = anthropic.Anthropic(
    api_key="YOUR_API_KEY",
    base_url="https://api.openveer.com"
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "解释量子计算的基本原理"}
    ]
)

print(message.content[0].text)

多轮对话

messages = [
    {"role": "user", "content": "什么是机器学习？"},
    {"role": "assistant", "content": "机器学习是人工智能的一个分支..."},
    {"role": "user", "content": "能举个实际应用的例子吗？"}
]

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=messages
)

使用系统提示词

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system="你是一位资深的Python开发专家，擅长代码审查和优化建议。",
    messages=[
        {"role": "user", "content": "如何优化这段代码？\n\n[代码]"}
    ]
)

流式响应

with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "写一篇关于AI的短文"}
    ]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

工具使用

tools = [
    {
        "name": "get_stock_price",
        "description": "获取股票的实时价格",
        "input_schema": {
            "type": "object",
            "properties": {
                "ticker": {
                    "type": "string",
                    "description": "股票代码，例如：AAPL"
                }
            },
            "required": ["ticker"]
        }
    }
]

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    tools=tools,
    messages=[
        {"role": "user", "content": "特斯拉的股价是多少？"}
    ]
)

# 处理工具调用
if message.stop_reason == "tool_use":
    tool_use = next(block for block in message.content if block.type == "tool_use")
    print(f"调用工具: {tool_use.name}")
    print(f"参数: {tool_use.input}")

视觉理解

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "source": {
                        "type": "url",
                        "url": "https://example.com/image.jpg"
                    }
                },
                {
                    "type": "text",
                    "text": "描述这张图片"
                }
            ]
        }
    ]
)

Base64图像

import base64

with open("image.jpg", "rb") as image_file:
    image_data = base64.b64encode(image_file.read()).decode("utf-8")

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "image",
                    "source": {
                        "type": "base64",
                        "media_type": "image/jpeg",
                        "data": image_data
                    }
                },
                {
                    "type": "text",
                    "text": "分析这张图片"
                }
            ]
        }
    ]
)

最佳实践

1. 提示词工程

清晰的角色定义：

system = """你是一位经验丰富的数据科学家，专长包括：
- 统计分析和数据可视化
- 机器学习模型开发
- Python和R编程
请提供专业、准确的建议。"""

结构化输出：

message = "请以JSON格式返回分析结果，包含summary、key_findings和recommendations字段。"

2. 错误处理

from anthropic import APIError, RateLimitError

try:
    message = client.messages.create(
        model="claude-sonnet-4-6",
        max_tokens=1024,
        messages=[{"role": "user", "content": "你好"}]
    )
except RateLimitError:
    print("速率限制，请稍后重试")
except APIError as e:
    print(f"API错误: {e}")

3. Token优化

# 使用更短的提示词
messages = [
    {"role": "user", "content": "总结要点：\n\n[长文本]"}
]

# 限制输出长度
message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=500,  # 限制输出
    messages=messages
)

4. 预填充响应

# 引导模型以特定格式回复
messages = [
    {"role": "user", "content": "列出5个Python最佳实践"},
    {"role": "assistant", "content": "以下是5个Python最佳实践：\n\n1."}
]

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=messages
)

流式响应处理

Python流式示例

import anthropic

client = anthropic.Anthropic(
    api_key="YOUR_API_KEY",
    base_url="https://api.openveer.com"
)

with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "写一个Python装饰器示例"}
    ]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

JavaScript流式示例

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

const client = new Anthropic({
  apiKey: process.env.API_KEY,
  baseURL: 'https://api.openveer.com'
});

const stream = await client.messages.stream({
  model: 'claude-sonnet-4-6',
  max_tokens: 1024,
  messages: [
    { role: 'user', content: '写一个React组件示例' }
  ]
});

for await (const chunk of stream) {
  if (chunk.type === 'content_block_delta' && 
      chunk.delta.type === 'text_delta') {
    process.stdout.write(chunk.delta.text);
  }
}

平台差异与对接注意

响应无包装

POST /v1/messages 成功时直接返回 Anthropic message 对象，没有 {code, data} 外层。官方 SDK、Claude Code、Cline 等才能 1:1 兼容。

错误格式（与官方唯一实质差异）

{
  "error": {
    "code": "model_not_found",
    "message": "... (request id: ...)",
    "param": "",
    "type": "apimart_error"
  }
}

相对 Anthropic 官方：顶层缺少 "type": "error"；error.type 固定为 apimart_error，而非 invalid_request_error 等语义化类型。

对接建议：不要依赖 error.type 做重试分支，改用 HTTP 状态码 + error.code：

报障请提供：error.message 末尾的 request id，以及响应头 x-oneapi-request-id。

流式 SSE

请求加 "stream": true。事件序列与官方一致：

message_start → content_block_start → ping → content_block_delta（多次）→ content_block_stop → message_delta → message_stop

⚠️ 流式与非流式的 usage 结构不同：message_delta.usage 通常只有 4 个 token 字段，没有 cache_creation、service_tier、inference_geo。请分开解析或全部设为可选。

未实现接口

POST /v1/messages/count_tokens 未实现，返回 404。官方 SDK 的 client.messages.count_tokens() 会失败。需要预估 token 时请在本地估算，或读取响应中的 usage.input_tokens。

必须忽略未知字段

本接口对上游透传，Anthropic 可能随时新增字段（如 stop_details、inference_geo、caller、output_tokens_details）。请勿开启严格 schema：

• Go：不要用 DisallowUnknownFields()
• Pydantic：不要 extra="forbid"
• TypeScript / Zod：用 .passthrough() 而非 .strict()

模型名建议

带 -thinking 后缀的同名模型是平台扩展别名。推荐用不带后缀的标准模型名 + 请求体 thinking 参数，便于迁移官方端点。

请求体其它字段与官方一致：model、messages、max_tokens（必填）、system、temperature、top_p、top_k、stop_sequences、stream、tools、tool_choice、thinking、metadata。语义以 Anthropic Messages API 为准。

注意事项

1. API 密钥安全：
   * 使用环境变量存储 API 密钥
   * 不要在代码中硬编码密钥
   * 定期轮换密钥

2. 速率限制：
   * 注意 API 的速率限制
   * 实现重试机制（按 HTTP 状态码）
   * 使用指数退避策略

3. Token 管理：
   * 监控 token 使用量（读 usage）
   * 优化提示词长度
   * 使用适当的 max_tokens 值
   * 开启 thinking 时 output_tokens 已含 thinking，勿重复计费

4. 模型选择：
   * Opus: 复杂任务、需要深度思考
   * Sonnet: 平衡性能和成本
   * Haiku: 快速响应、简单任务

5. 内容解析：
   * 遍历 content 取 type == "text"，不要写死 content[0].text
   * 模型若返回 Markdown 代码块包裹的 JSON，属模型输出而非接口包装（见下方 FAQ）

6. 内容过滤：
   * 验证用户输入
   * 过滤敏感信息
   * 实现内容审核机制

FAQ

响应里 content 的 text 是 ` ```json

这不是接口结构问题。text 字段装的是模型生成的原始内容：模型判断你想要 JSON，就用 Markdown 代码块包起来了。接口不会也不应该改写模型输出。

想拿到干净的结构化数据，有三种正确做法（推荐程度从高到低）：

1. 用 tools 强制结构化输出——最可靠，input 字段直接就是解析好的对象：

{
  "tools": [{
    "name": "emit_result",
    "input_schema": {
      "type": "object",
      "properties": { "answer": { "type": "string" } }
    }
  }],
  "tool_choice": { "type": "tool", "name": "emit_result" }
}

2. prefill 助手消息，让模型从 { 接着写：

{
  "messages": [
    { "role": "user", "content": "..." },
    { "role": "assistant", "content": "{" }
  ]
}

3. 在 system prompt 里明确要求「只输出 JSON，不要加 Markdown 代码块」。

不推荐用正则去剥 code fence——模型偶尔不加围栏时会解析失败。