Anthropic Messages

Claude modellerine extended thinking, prompt caching, tool use, web search/fetch, streaming ve effort control ile erişmek için Anthropic Messages API’yi CometAPI üzerinden kullanın.

POST /v1/messages

Python

Basic

import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key="<COMETAPI_KEY>",
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system="You are a helpful assistant.",
    messages=[
        {"role": "user", "content": "Hello, world"}
    ],
)

print(message.content[0].text)

200

{
  "id": "<string>",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "<string>",
      "thinking": "<string>",
      "signature": "<string>",
      "id": "<string>",
      "name": "<string>",
      "input": {}
    }
  ],
  "model": "<string>",
  "stop_reason": "end_turn",
  "stop_sequence": "<string>",
  "usage": {
    "input_tokens": 123,
    "output_tokens": 123,
    "cache_creation_input_tokens": 123,
    "cache_read_input_tokens": 123,
    "cache_creation": {
      "ephemeral_5m_input_tokens": 123,
      "ephemeral_1h_input_tokens": 123
    }
  }
}

​

Genel Bakış

CometAPI, Anthropic Messages API’yi yerel olarak destekler ve size tüm Anthropic’e özgü özelliklerle birlikte Claude modellerine doğrudan erişim sağlar. Extended thinking, prompt caching ve effort control gibi yalnızca Claude’a özel yetenekler için bu endpoint’i kullanın.

​

Hızlı Başlangıç

Resmi Anthropic SDK’sını kullanın — sadece base URL’yi CometAPI olarak ayarlayın:

Python

import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key="<COMETAPI_KEY>",
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}],
)
print(message.content[0].text)

Kimlik doğrulama için hem x-api-key hem de Authorization: Bearer başlıkları desteklenir. Resmi Anthropic SDK’ları varsayılan olarak x-api-key kullanır.

​

Extended Thinking

thinking parametresiyle Claude’un adım adım akıl yürütmesini etkinleştirin. Yanıt, nihai yanıttan önce Claude’un dahili akıl yürütmesini gösteren thinking içerik bloklarını içerir.

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=16000,
    thinking={
        "type": "enabled",
        "budget_tokens": 10000,
    },
    messages=[
        {"role": "user", "content": "Prove that there are infinitely many primes."}
    ],
)

for block in message.content:
    if block.type == "thinking":
        print(f"Thinking: {block.thinking[:200]}...")
    elif block.type == "text":
        print(f"Answer: {block.text}")

Thinking için minimum budget_tokens değeri 1,024 olmalıdır. Thinking tokens, max_tokens sınırınıza dahil edilir — hem thinking hem de yanıtı karşılayacak kadar yüksek bir max_tokens değeri ayarlayın.

---

​

Prompt Caching

Sonraki isteklerde gecikmeyi ve maliyeti azaltmak için büyük system prompt’ları veya konuşma öneklerini önbelleğe alın. Önbelleğe alınması gereken içerik bloklarına cache_control ekleyin:

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system=[
        {
            "type": "text",
            "text": "You are an expert code reviewer. [Long detailed instructions...]",
            "cache_control": {"type": "ephemeral"},
        }
    ],
    messages=[{"role": "user", "content": "Review this code..."}],
)

Önbellek kullanımı, yanıttaki usage alanında raporlanır:

• cache_creation_input_tokens — önbelleğe yazılan tokens (daha yüksek bir ücretle faturalandırılır)
• cache_read_input_tokens — önbellekten okunan tokens (indirimli bir ücretle faturalandırılır)

Prompt caching, önbelleğe alınan içerik bloğunda minimum 1,024 tokens gerektirir. Bundan daha kısa içerikler önbelleğe alınmaz.

---

​

Streaming

stream: true ayarını yaparak yanıtları Server-Sent Events (SSE) kullanarak akış halinde alın. Event’ler şu sırayla gelir:

1. message_start — mesaj meta verilerini ve başlangıç usage bilgisini içerir
2. content_block_start — her content bloğunun başlangıcını işaret eder
3. content_block_delta — artımlı metin parçaları (text_delta)
4. content_block_stop — her content bloğunun sonunu işaret eder
5. message_delta — son stop_reason ve tam usage
6. message_stop — akışın sonunu bildirir

with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=256,
    messages=[{"role": "user", "content": "Hello"}],
) as stream:
    for text in stream.text_stream:
        print(text, end="")

---

​

Effort Control

Claude’un bir yanıt oluştururken ne kadar çaba harcayacağını output_config.effort ile kontrol edin:

message = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=4096,
    messages=[
        {"role": "user", "content": "Summarize this briefly."}
    ],
    output_config={"effort": "low"},  # "low", "medium", or "high"
)

---

​

Server Tools

Claude, Anthropic altyapısında çalışan sunucu taraflı araçları destekler:

Web Fetch

URL’lerden içerik getirip analiz edin:

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Analyze the content at https://arxiv.org/abs/1512.03385"}
    ],
    tools=[
        {"type": "web_fetch_20250910", "name": "web_fetch", "max_uses": 5}
    ],
)

Web Search

Gerçek zamanlı bilgiler için web’de arama yapın:

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "What are the latest developments in AI?"}
    ],
    tools=[
        {"type": "web_search_20250305", "name": "web_search", "max_uses": 5}
    ],
)

---

​

Yanıt Örneği

CometAPI’nin Anthropic endpoint’inden tipik bir yanıt:

{
  "id": "msg_bdrk_01UjHdmSztrL7QYYm7CKBDFB",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "Hello!"
    }
  ],
  "model": "claude-sonnet-4-6",
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": {
    "input_tokens": 19,
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0,
    "cache_creation": {
      "ephemeral_5m_input_tokens": 0,
      "ephemeral_1h_input_tokens": 0
    },
    "output_tokens": 4
  }
}

---

​

OpenAI-Compatible Endpoint’ten Temel Farklar

Özellik	Anthropic Messages (/v1/messages)	OpenAI-Compatible (/v1/chat/completions)
Extended thinking	budget_tokens ile thinking parametresi	Kullanılamaz
Prompt caching	content bloklarında cache_control	Kullanılamaz
Effort control	output_config.effort	Kullanılamaz
Web fetch/search	Server tools (web_fetch, web_search)	Kullanılamaz
Auth header	x-api-key veya Bearer	Yalnızca Bearer
Response format	Anthropic formatı (content blokları)	OpenAI formatı (choices, message)
Models	Yalnızca Claude	Çok sağlayıcılı (GPT, Claude, Gemini, vb.)

Yetkilendirmeler

x-api-key · string · header · gerekli

Your CometAPI key passed via the x-api-key header. Authorization: Bearer <key> is also supported.

Başlıklar

anthropic-version · string · varsayılan: 2023-06-01

The Anthropic API version to use. Defaults to 2023-06-01.

Örnek:

"2023-06-01"

anthropic-beta · string

Comma-separated list of beta features to enable. Examples: max-tokens-3-5-sonnet-2024-07-15, pdfs-2024-09-25, output-128k-2025-02-19.

Gövde

application/json

model · string · gerekli

The Claude model to use. See the Models page for current Claude model IDs.

Örnek:

"claude-sonnet-4-6"

messages · object[] · gerekli

The conversation messages. Must alternate between user and assistant roles. Each message's content can be a string or an array of content blocks (text, image, document, tool_use, tool_result). There is a limit of 100,000 messages per request.

Show child attributes

max_tokens · integer · gerekli

The maximum number of tokens to generate. The model may stop before reaching this limit. When using thinking, the thinking tokens count towards this limit.

Gerekli aralık: x >= 1

Örnek:

1024

system · string object[]

System prompt providing context and instructions to Claude. Can be a plain string or an array of content blocks (useful for prompt caching).

temperature · number · varsayılan: 1

Controls randomness in the response. Range: 0.0–1.0. Use lower values for analytical tasks and higher values for creative tasks. Defaults to 1.0.

Gerekli aralık: 0 <= x <= 1

top_p · number

Nucleus sampling threshold. Only tokens with cumulative probability up to this value are considered. Range: 0.0–1.0. Use either temperature or top_p, not both.

Gerekli aralık: 0 <= x <= 1

top_k · integer

Only sample from the top K most probable tokens. Recommended for advanced use cases only.

Gerekli aralık: x >= 0

stream · boolean · varsayılan: false

If true, stream the response incrementally using Server-Sent Events (SSE). Events include message_start, content_block_start, content_block_delta, content_block_stop, message_delta, and message_stop.

stop_sequences · string[]

Custom strings that cause the model to stop generating when encountered. The stop sequence is not included in the response.

thinking · object

Enable extended thinking — Claude's step-by-step reasoning process. When enabled, the response includes thinking content blocks before the answer. Requires a minimum budget_tokens of 1,024.

Show child attributes

tools · object[]

Tools the model may use. Supports client-defined functions, web search (web_search_20250305), web fetch (web_fetch_20250910), code execution (code_execution_20250522), and more.

Show child attributes

tool_choice · object

Controls how the model uses tools.

Show child attributes

metadata · object

Request metadata for tracking and analytics.

Show child attributes

output_config · object

Configuration for output behavior.

Show child attributes

service_tier · enum<string>

The service tier to use. auto tries priority capacity first, standard_only uses only standard capacity.

Mevcut seçenekler: auto, standard_only

Yanıt

200 - application/json

Successful response. When stream is true, the response is a stream of SSE events.

id · string

Unique identifier for this message (e.g., msg_01XFDUDYJgAACzvnptvVoYEL).

type · enum<string>

Always message.

Mevcut seçenekler: message

role · enum<string>

Always assistant.

Mevcut seçenekler: assistant

content · object[]

The response content blocks. May include text, thinking, tool_use, and other block types.

Show child attributes

model · string

The specific model version that generated this response (e.g., claude-sonnet-4-6).

stop_reason · enum<string>

Why the model stopped generating.

Mevcut seçenekler: end_turn, max_tokens, stop_sequence, tool_use, pause_turn

stop_sequence · string | null

The stop sequence that caused the model to stop, if applicable.

usage · object

Token usage statistics.

Show child attributes