Messages
POST /v1/messages
Use the Messages surface when an agent or SDK expects the Anthropic request, response and streaming event shapes. The same AIx key, wallet, routing controls and metering apply.
For OpenAI-compatible clients, use /v1/chat/completions.
Choose one surface per client; both reach the same AIx model catalog.
Request body
| Field | Type | Required | Description |
|---|---|---|---|
model | string | required | Canonical AIx model id, or a supported Claude-style model alias. |
messages | array | required | Conversation turns with user or assistant roles. Content may be a string or an array of text, image, tool-use and tool-result blocks. |
system | string | array | optional | System instruction as plain text or an array of text blocks. |
max_tokens | integer | default: 256 | Maximum output tokens. |
temperature | number | default: 0.7 | Sampling temperature. |
top_p | number | default: 1.0 | Nucleus sampling. |
stop_sequences | array | optional | Custom stop sequences. |
stream | boolean | default: false | Return Anthropic-style server-sent events. |
tools | array | optional | Tool definitions using name, description and input_schema. |
tool_choice | object | optional | auto, any, none or one named tool. |
Example
curl https://api.aix.theaimart.co/v1/messages \
-H "Authorization: Bearer $AIX_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "anthropic/claude-sonnet-4-6",
"max_tokens": 512,
"messages": [{"role":"user","content":"Explain this API in one sentence."}]
}'import os
import anthropic
client = anthropic.Anthropic(
api_key=os.environ["AIX_KEY"],
base_url="https://api.aix.theaimart.co",
)
message = client.messages.create(
model="anthropic/claude-sonnet-4-6",
max_tokens=512,
messages=[{"role": "user", "content": "Explain this API in one sentence."}],
)import Anthropic from "@anthropic-ai/sdk";
const client = new Anthropic({
apiKey: process.env.AIX_KEY,
baseURL: "https://api.aix.theaimart.co",
});
const message = await client.messages.create({
model: "anthropic/claude-sonnet-4-6",
max_tokens: 512,
messages: [{ role: "user", content: "Explain this API in one sentence." }],
});Response — message
{
"id": "msg_…",
"type": "message",
"role": "assistant",
"model": "anthropic/claude-sonnet-4-6",
"content": [{ "type": "text", "text": "…" }],
"stop_reason": "end_turn",
"stop_sequence": null,
"usage": { "input_tokens": 18, "output_tokens": 24 }
}
Tool calls are returned as tool_use content blocks. Send their results back as
tool_result blocks with the matching tool_use_id.
Streaming
With stream: true, the response is text/event-stream. Events follow this lifecycle:
message_start → content_block_start → content_block_delta
→ content_block_stop → message_delta → message_stop
Text deltas use text_delta; tool arguments use input_json_delta. A ping event may appear
and should be ignored by clients that do not need heartbeats.
Errors
This endpoint uses the Messages error envelope:
{
"type": "error",
"error": { "type": "rate_limit_error", "message": "…" }
}
Retry 429, 500 and 503 with exponential backoff and jitter. Do not retry validation,
authentication or billing errors without changing the request or account state.