Claude Messages API

Send prompts to Claude Opus 4.7, Sonnet 4.6 via the Messages API. OpenAI-compatible. Code examples, request/response format, streaming support.

Updated: July 25, 2026.

Endpoint and headers

Claude API Tech supports the Anthropic-compatible Messages API. Send a POST request to the endpoint below and keep the API key on your server.

POST

https://api.llm-gate.tech/v1/messages

Header	Value
x-api-key	Your Claude API Tech key
anthropic-version	API version, such as 2023-06-01
content-type	application/json

OpenAI-compatible API

For clients built around the OpenAI format, use the compatible endpoint: https://api.llm-gate.tech/v1/chat/completions

Your first Claude API request

A minimal request includes model, max_tokens, and a messages array. The tabs show the same request in cURL, Python, and TypeScript.

request.sh

curl https://api.llm-gate.tech/v1/messages \  --header "x-api-key: $CLAUDE_API_KEY" \  --header "anthropic-version: 2023-06-01" \  --header "content-type: application/json" \  --data '{    "model": "claude-sonnet-4-6",    "max_tokens": 512,    "messages": [      {"role": "user", "content": "Explain SSE in one paragraph"}    ]  }'

Request body format

Field	Required	Purpose
model	Yes	The exact ID of an available Claude model.
max_tokens	Yes	Maximum number of tokens to generate.
messages	Yes	Conversation history with user and assistant messages.
system	No	Instructions that apply from the start of the request.
stream	No	Set true to receive a streaming SSE response.
tools	No	Tools that the model is allowed to call.

Streaming

Set stream: true in the request body to receive Server-Sent Events. See the Streaming API page for details.

Message roles and format

The Messages API is stateless. To continue a conversation, include the required history again in every request.

user — an instruction or message from the user.
assistant — a previous Claude response or stored conversation turn.
system — use the top-level system field for instructions that apply from the beginning.
content — can be a string or an array of text, image, tool_use, and tool_result blocks.

The Messages API is stateless

The API does not remember a previous call. Store history in your application, remove unnecessary old turns, and keep the full array within the model's context window.

Messages API response format

A successful response contains a content array, a stop_reason, and actual token usage under usage.

JSON

{
  "id": "msg_...",
  "type": "message",
  "role": "assistant",
  "content": [
    {"type": "text", "text": "SSE is a one-way HTTP stream..."}
  ],
  "model": "claude-sonnet-4-6",
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 18,
    "output_tokens": 42
  }
}

content — response blocks; generated text is usually in a type: text block.
stop_reason — why generation ended, such as end_turn, max_tokens, or tool_use.
usage.input_tokens and usage.output_tokens — the tokens consumed by the request.

Streaming API →