Anthropic Messages

Create a message using the Anthropic Claude API format. Text input only; supports both streaming (SSE) and non-streaming responses.

POSThttps://api.longcat.chat/anthropic/v1/messages

Authorizations

Authorizationstringheaderrequired

Authentication key, passed in the request header as Bearer YOUR_API_KEY.

default: false

Bodyapplication/json

modelstringrequired

Model identifier. Currently supports LongCat-2.0-Preview.

default: false

messagesarrayrequired

Array of message objects. Text input only.

default: false

rolestringrequired

Author role, one of user · assistant. System messages are passed separately via the system parameter.

default: false

contentstringrequired

Message content as a plain-text string.

default: false

max_tokensintegeroptional

Maximum number of tokens to generate.

default: false

streambooleanoptional

Whether to stream the response as SSE.

default: false

temperaturenumberoptional

Sampling temperature, range 0 ~ 1.

default: false

top_pnumberoptional

Nucleus sampling parameter.

default: false

systemstringoptional

System message used to set context.

default: false

thinkingobjectoptional

Whether to enable thinking. {"type":"enabled"} turns thinking on, {"type":"disabled"} turns it off.

default: false

Request

cURL

Basic

bash

curl --location --request POST 'https://api.longcat.chat/anthropic/v1/messages' \
  --header "Authorization: Bearer $LONGCAT_API_KEY" \
  --header "Content-Type: application/json" \
  --data-raw '{
    "model": "LongCat-2.0-Preview",
    "max_tokens": 1000,
    "messages": [{ "role": "user", "content": "Hello, LongCat" }],
    "thinking": {
      "type": "enabled"
    }
  }'

Response

JSON

json

{
  "id": "msg_123",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "Hello! How can I help you today?",
      "thinking": "\nThe user greeted me in English, so I should respond in English to maintain consistency. The greeting is simple and friendly, so a warm and polite reply is appropriate. \n\nI recall that LongCat is developed by Meituan, so I should introduce myself naturally without sounding robotic. The response should acknowledge the greeting and express willingness to assist. \n\nI'll keep it concise but friendly, ending with an open question to encourage further interaction."
    }
  ],
  "model": "LongCat-2.0-Preview",
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": { "input_tokens": 10, "output_tokens": 8 }
}