Anthropic Messages API

Use /v1/messages to call any model with the Anthropic SDK. The proxy translates requests to OpenAI chat-completions and back.

POST /v1/messages accepts the Anthropic Messages API request shape. The proxy translates it to OpenAI chat-completions, routes it through your enabled upstreams, and translates the response (including SSE) back into Anthropic's format. This lets you use the Anthropic SDK to reach models hosted on any provider — not just Anthropic.

Authentication

Authorization: Bearer sk-rc-...

Use your Ada AI proxy key (an sk-rc-… value), not an Anthropic API key.

Request

curl https://api.ada.ai/v1/messages \
  -H "Authorization: Bearer sk-rc-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-4.7",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

The body follows Anthropic's schema. Notable fields:

Field	Required	Notes
`model`	yes	Must match a model ID from one of your enabled upstreams. The `model` is not restricted to Anthropic models — any model you've enabled works.
`max_tokens`	yes	Anthropic requires this field.
`messages`	yes	Anthropic message array. `content` may be a string or an array of content blocks.
`system`	no	Top-level system prompt (Anthropic convention).
`stream`	no	`true` returns Anthropic-format SSE.
`tools`	no	Anthropic-style tool definitions.

Text, image, tool-use, and tool-result content blocks are all supported and translated to their OpenAI equivalents.

Response

Anthropic-shaped:

{
  "id": "msg_...",
  "type": "message",
  "role": "assistant",
  "model": "glm-4.7",
  "content": [{ "type": "text", "text": "Hello! How can I help?" }],
  "stop_reason": "end_turn",
  "usage": { "input_tokens": 8, "output_tokens": 7 }
}

Streaming

Set "stream": true to receive Anthropic-format SSE. The proxy synthesizes it from the upstream OpenAI chunks, so you see the standard Anthropic events:

event: message_start
event: content_block_start
event: content_block_delta
event: content_block_stop
event: message_delta
event: message_stop

Using the Anthropic SDK

Point the SDK's baseURL at Ada AI and pass your proxy key as apiKey:

import anthropic

client = anthropic.Anthropic(
    api_key="sk-rc-...",
    base_url="https://api.ada.ai",
)

message = client.messages.create(
    model="glm-4.7",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}],
)
print(message.content[0].text)

base_url has no /v1

The Anthropic SDK appends /v1/messages itself, so set base_url (Python) or baseURL (JS) to https://api.ada.ai — without a trailing /v1. The OpenAI SDK, by contrast, takes https://api.ada.ai/v1.

When to use which endpoint

You have...	Use
OpenAI client code already	`/v1/chat/completions`
Anthropic client code already	`/v1/messages` (this page)
A free choice	Either — same routing, same upstreams, same key

Both endpoints resolve against the same scoped upstreams for your key, so the model you request reaches the same provider either way.