Docs
API reference
Messages

Messages

POSThttps://api.airoutly.com/v1/messages

Anthropic's native API for Claude models, used by the Anthropic SDKs and Claude Code. Set the SDK's Base URL to https://api.airoutly.com, without /v1.

Authentication

Send x-api-key: <key> (what the Anthropic SDKs do) or Authorization: Bearer <key>. The anthropic-version header is accepted as usual.

Request body

Parameter	Description
`model`stringrequired	A model that lists Msg on the models page, such as `claude-sonnet-4-5`.
`max_tokens`integerrequired	Upper limit on generated tokens.
`messages`arrayrequired	Alternating `user` and `assistant` turns; `content` can include text, images and tool results.
`system`string \| array	System prompt. Use the array form to add `cache_control`.
`stream`boolean	Stream the reply as server-sent events.
`tools`array	Tools the model may call.
`thinking`object	Extended thinking on models that support it, for example `{"type": "enabled", "budget_tokens": 4096}`.

Example request

curl https://api.airoutly.com/v1/messages \
  -H "x-api-key: $AIROUTLY_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

import os
from anthropic import Anthropic

client = Anthropic(base_url="https://api.airoutly.com", api_key=os.environ["AIROUTLY_API_KEY"])
msg = client.messages.create(
    model="claude-sonnet-4-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
)
print(msg.content[0].text)

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

const client = new Anthropic({ baseURL: "https://api.airoutly.com", apiKey: process.env.AIROUTLY_API_KEY });
const msg = await client.messages.create({
  model: "claude-sonnet-4-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello" }],
});

Prompt caching

Mark long, repeated content with cache_control. Writing to the cache is billed at the cache-write price, and later hits at the much lower cache-read price, both shown on each model’s page.

Python

msg = client.messages.create(
    model="claude-sonnet-4-5",
    max_tokens=1024,
    system=[{
        "type": "text",
        "text": LONG_SYSTEM_PROMPT,
        "cache_control": {"type": "ephemeral"},
    }],
    messages=[{"role": "user", "content": "Summarise section 3."}],
)
print(msg.usage.cache_creation_input_tokens, msg.usage.cache_read_input_tokens)

Response

200 OK

{
  "id": "msg_123",
  "type": "message",
  "role": "assistant",
  "model": "claude-sonnet-4-5",
  "content": [{ "type": "text", "text": "Hello! How can I help?" }],
  "stop_reason": "end_turn",
  "usage": { "input_tokens": 9, "output_tokens": 8, "cache_creation_input_tokens": 0, "cache_read_input_tokens": 0 }
}

Token countingThe /v1/messages/count_tokens endpoint isn’t available. Read token counts from usage in the response instead.