Create embeddings (OpenAI-compatible)
OpenAI's embeddings API. input is a string or an array of strings (token-ID arrays are refused with token_input_unsupported); dimensions truncates the vectors (Matryoshka), within the model's minDimensions and dimensions; encoding_format is float (the default) or base64. user is accepted and ignored. Errors use OpenAI's body, with the platform's stable code in error.code.
https://inference-api.ezghcloud.com/v1/organizations/{orgId}/projects/{projectId}/openai/v1/embeddingsAuthorization
Path parameters
Request bodyJSONrequired
Returns
200
One embedding per input, in input order.
400
invalid_request_error (invalid_request, or the product's own code): the request isn't valid; param names the field.
401
authentication_error (unauthenticated): no bearer credential, or one that doesn't check out. The session cookie isn't read.
403
permission_error (access_denied): the caller's policies don't allow the action.
404
not_found_error (not_found, or the product's own code): unknown, malformed, or in an organization the caller isn't in.
409
The organization is being deleted (organization_deletion_pending). Nothing was computed or metered.
413
invalid_request_error (payload_too_large): the body is over the endpoint's limit.
429
rate_limit_error (too_many_requests), with Retry-After; or insufficient_quota (insufficient_quota), a monthly limit, without one and with x-should-retry: false.
500
api_error (internal_error): a bug; the details are in the logs, under the request ID.
502
The model backend failed or returned an incomplete result (inference_failed). Nothing was metered.
503
api_error (unavailable): a dependency the request needs is down. Retry after Retry-After.