Bypass Anthropic Messages

Native protocol and authentication

Use an XHub public model name and virtual key with the matching Bypass deployment. XHub maps the model and injects deployment credentials, preserves native fields and enforces access and billing. Bypass does not run the standard pre-call text guardrail chain. Send Authorization: Bearer $XHUB_API_KEY. Model availability depends on configured deployments. x-api-key: $XHUB_API_KEY is also accepted. Send anthropic-version: 2023-06-01. Do not send provider credentials or conflicting keys.

Overview

Anthropic Messages

Behavior

Use a Messages deployment. Native clients send anthropic-version: 2023-06-01. XHub also accepts its key in x-api-key. Tool use/results are content blocks.

Endpoint and authentication

POST /bypass/anthropic/v1/messages

Use an XHub API key: Authorization: Bearer . Do not use upstream provider credentials as gateway keys.

Use the gateway root Base URL shown in the playground, without appending /v1; send requests to the full endpoint path above.

Request example

curl "$XHUB_BASE_URL/bypass/anthropic/v1/messages" \
  -H "Authorization: Bearer $XHUB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"YOUR_MODEL_NAME","max_tokens":256,"messages":[{"role":"user","content":"Hello, introduce yourself."}]}'

Request parameters

Field	Type	Description
model	string · required	Public model name; deployment must support this protocol.
max_tokens	integer · required	Maximum output tokens.
messages	object[] · required	User/assistant content blocks.
system	string or array	Top-level system instructions.
tools / stream	array / boolean	Anthropic tool schema / event stream.

Success response

{
  "id": "msg-example",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "Hello!"
    }
  ],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 10,
    "output_tokens": 5
  }
}

Field	Type
id	string
type	string
role	string
content	object[]
content[].type	string
content[].text	string
stop_reason	string
usage	object
usage.input_tokens	number
usage.output_tokens	number

Errors and troubleshooting

HTTP	Meaning
400	Invalid request. Check JSON, parameters and query.
401	Missing or invalid XHub credential.
403	Permission denied. Check model scope or policy.
429	Limit exceeded. Inspect response details and Retry-After.