Claude Messages (Anthropic native)

Call Claude with Anthropic's official Messages API, with full native support for extended thinking, prompt caching, tool use and more.

Endpoint

POST https://api.onefast.ai/v1/messages
x-api-key: sk-xxx
anthropic-version: 2023-06-01
Content-Type: application/json

The native endpoint is the recommended way to call Claude: requests and responses are identical to Anthropic's official API, so you get extended thinking, prompt caching, anthropic-beta and every other capability. With the official Anthropic SDKs you only change base_url.

Set the Anthropic SDK's base_url to the root URL https://api.onefast.ai — without /v1 (the SDK appends /v1/messages itself).

Basic request

cURL

curl https://api.onefast.ai/v1/messages \
  -H "x-api-key: $API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello, Claude"}]
  }'

Python

# pip install anthropic
import anthropic

client = anthropic.Anthropic(api_key="sk-xxx", base_url="https://api.onefast.ai")
message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello, Claude"}],
)
print(message.content[0].text)

Node.js

// npm install @anthropic-ai/sdk
import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({ apiKey: "sk-xxx", baseURL: "https://api.onefast.ai" });
const message = await client.messages.create({
  model: "claude-sonnet-4-6",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello, Claude" }],
});
console.log(message.content[0].text);

Example response:

{
  "id": "msg_...",
  "type": "message",
  "role": "assistant",
  "model": "claude-sonnet-4-6",
  "content": [{ "type": "text", "text": "Hello! How can I help you today?" }],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 10,
    "output_tokens": 12,
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0
  }
}

Headers

Streaming

Python

with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a short poem about autumn"}],
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

The event sequence matches the official API: message_start → content_block_delta … → message_delta (with final usage) → message_stop.

Extended thinking

Python

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=16000,
    thinking={"type": "enabled", "budget_tokens": 8000},
    messages=[{"role": "user", "content": "Prove that the square root of 2 is irrational"}],
)
for block in message.content:
    print(block.type, ":", getattr(block, "thinking", None) or getattr(block, "text", ""))

Thinking tokens are billed as output tokens. See Tools & reasoning for more.

Counting tokens

Estimate input tokens before sending — not billed:

cURL

curl https://api.onefast.ai/v1/messages/count_tokens \
  -H "x-api-key: $API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-4-6", "messages": [{"role": "user", "content": "Hello, Claude"}]}'

Returns {"input_tokens": 9}. This is a local estimate by the gateway and may differ slightly from actual billing.

Calling non-Claude models through this endpoint

/v1/messages can also call GPT, DeepSeek and other OpenAI-protocol models; the gateway converts the format automatically, which makes switching models in tools like Claude Code easy. Native Gemini routes can't be called through this endpoint yet.