Sign in
Private Endpoints

Create a private endpoint

Create a private endpoint as a draft. Drafts are not routable: validate one with POST /private-endpoints/{id}/validate, then activate it with POST /private-endpoints/{id}/activate. Pass activate to do all three in one call; if validation fails the draft is kept and returned with the failed checks. Send an Idempotency-Key header to make retries safe: a repeated key with the same request returns the endpoint the first request created; reusing it with different fields returns 422 idempotency_key_reused. Management key required.

post/private-endpoints

Headers

idempotency-keystring

Retry-safe create: a repeated create with the same key from the same organization returns the endpoint the first request created instead of creating another. The endpoint is returned as it is now. Reusing a key with a different request body (model, provider, base URL, upstream model ID, declared ZDR or region, or pricing) is rejected with 422 idempotency_key_reused. activate is not compared, and later edits to the endpoint do not affect the comparison.

Example:wayfair-gpt-4o-eastus-2026-09

Retry-safe create: a repeated create with the same key from the same organization returns the endpoint the first request created instead of creating another. The endpoint is returned as it is now. Reusing a key with a different request body (model, provider, base URL, upstream model ID, declared ZDR or region, or pricing) is rejected with 422 idempotency_key_reused. activate is not compared, and later edits to the endpoint do not affect the comparison.

Request body

base_urlstring

HTTPS base URL of your deployment. Required unless the provider derives its URL from the BYOK credential (Azure, Amazon Bedrock, Google Vertex).

declared_region'global' | 'europe' | 'us' | 'null' nullable

Attest where this deployment processes data.

declared_zdrboolean nullable

Attest that this deployment retains no prompt or completion data.

model_permaslugstring required

Permanent slug of the model this endpoint serves.

provider_slugstring required

Slug of the upstream provider.

upstream_model_idstring required

Model or deployment identifier sent to the upstream provider.

Example request

{
  "activate": {
    "workspace_id": "550e8400-e29b-41d4-a716-446655440000"
  },
  "base_url": "https://contoso.openai.azure.com",
  "declared_region": "us",
  "model_permaslug": "openai/gpt-4o-2024-08-06",
  "pricing": {
    "completion": "0.00001",
    "prompt": "0.0000025"
  },
  "provider_slug": "azure",
  "upstream_model_id": "gpt-4o-prod"
}

Response

The created endpoint: a draft, or active when activate was passed

Example response

{
  "data": {
    "created_at": "2026-09-24T10:30:00Z",
    "id": "5b1c4c4e-7d0a-4a8e-9f3a-2d6c1b0e8a11",
    "model_name": "OpenAI: GPT-4o",
    "model_permaslug": "openai/gpt-4o-2024-08-06",
    "model_slug": "openai/gpt-4o",
    "provider_name": "Azure",
    "status": "active",
    "declared_region": "us",
    "declared_zdr": true
  }
}

Changes

Changed in 1 of the 339 revisions of this API.1