Anthropic Messages - CometAPI Documentation

POST

messages

import os
import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key=os.environ["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)

{
  "id": "<string>",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "text": "<string>",
      "thinking": "<string>",
      "signature": "<string>",
      "id": "<string>",
      "name": "<string>",
      "input": {}
    }
  ],
  "model": "<string>",
  "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
    }
  }
}

CometAPI supports the Anthropic Messages API natively, giving you direct access to Claude models with all Anthropic-specific features. Use this endpoint for Claude-exclusive capabilities like extended thinking, prompt caching, and effort control.

Both x-api-key and Authorization: Bearer headers are supported for authentication. The official Anthropic SDKs use x-api-key by default.

Quick start

To use the official Anthropic SDK with CometAPI, set the base URL:

import os
import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key=os.environ["COMETAPI_KEY"],
)

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

Enable extended thinking

Enable Claude’s step-by-step reasoning with the thinking parameter. The response includes thinking content blocks showing Claude’s internal reasoning before the final answer.

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 requires a minimum budget_tokens of 1,024. Thinking tokens count towards your max_tokens limit — set max_tokens high enough to accommodate both thinking and the response.

Cache prompts

To reduce latency and cost on subsequent requests, cache large system prompts or conversation prefixes. Add cache_control to content blocks that should be cached:

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..."}],
)

Cache usage is reported in the response usage field:

cache_creation_input_tokens — tokens written to cache (billed at a higher rate)
cache_read_input_tokens — tokens read from cache (billed at a reduced rate)

Prompt caching requires a minimum of 1,024 tokens in the cached content block. Content shorter than this will not be cached.

Stream responses

To stream responses using Server-Sent Events (SSE), set stream: true. Events arrive in this order:

message_start — contains the message metadata and initial usage
content_block_start — marks the beginning of each content block
content_block_delta — incremental text chunks (text_delta)
content_block_stop — marks the end of each content block
message_delta — final stop_reason and complete usage
message_stop — signals the end of the stream

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="")

Control effort

To control how much effort Claude puts into generating a response, use output_config.effort:

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"
)

Use server tools

Claude supports server-side tools that run on Anthropic’s infrastructure:

Web Fetch
Web Search

Fetch and analyze content from URLs:

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}
    ],
)

Search the web for real-time information:

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}
    ],
)

Response example

A typical response from CometAPI’s Anthropic endpoint:

{
  "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
  }
}

Compare with OpenAI-compatible endpoint

Feature	Anthropic Messages (`/v1/messages`)	OpenAI-Compatible (`/v1/chat/completions`)
Extended thinking	`thinking` parameter with `budget_tokens`	Not available
Prompt caching	`cache_control` on content blocks	Not available
Effort control	`output_config.effort`	Not available
Web fetch/search	Server tools (`web_fetch`, `web_search`)	Not available
Auth header	`x-api-key` or `Bearer`	`Bearer` only
Response format	Anthropic format (`content` blocks)	OpenAI format (`choices`, `message`)
Models	Claude only	Multi-provider (GPT, Claude, Gemini, etc.)

Authorizations

x-api-key

string

header

required

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

Headers

anthropic-version

string

default:2023-06-01

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

Example:

"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.

Body

application/json

model

string

required

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

Example:

"claude-sonnet-4-6"

messages

object[]

required

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

required

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.

Required range: x >= 1

Example:

1024

system

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

default: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.

Required range: 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.

Required range: 0 <= x <= 1

top_k

integer

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

Required range: x >= 0

stream

boolean

default: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.

Available options:

auto,

standard_only

Response

200 - application/json

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

string

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

type

enum<string>

Always message.

Available options:

message

role

enum<string>

Always assistant.

Available options:

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.

Available options:

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

Responses

Gemini Generating Content

import os
import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key=os.environ["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)

{
  "id": "<string>",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "text": "<string>",
      "thinking": "<string>",
      "signature": "<string>",
      "id": "<string>",
      "name": "<string>",
      "input": {}
    }
  ],
  "model": "<string>",
  "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
    }
  }
}

​Quick start

​Enable extended thinking

​Cache prompts

​Stream responses

​Control effort

​Use server tools

​Response example

​Compare with OpenAI-compatible endpoint

Authorizations

Headers

Body

Response

Quick start

Enable extended thinking

Cache prompts

Stream responses

Control effort

Use server tools

Response example

Compare with OpenAI-compatible endpoint