Completions

POST

/v1/completions

curl --request POST \
  --url http://localhost:8000/v1/completions \
  --header 'Content-Type: application/json' \
  --data '{}'

• Local mock mode (make demo)
• Your gateway (the public demo's host by default)

The legacy completions API: each prompt runs the chat path as one user message (_serve), with the same checks, task resolution, cascade, quota and storage as /v1/chat/completions; n, stream, stop, max_tokens and the rest are the chat fields of the same name. A batch of prompts needs that many requests of the daily quota, checked before the first runs; a batch that stops after its first prompt is an error that names the prompts already answered and carries X-Should-Retry: false (_batch_stopped). A suffix, echo or best_of that would change the answer, and token-array prompts, are a 400.

Parameters

Header Parameters

x-ratchet-task

Any of:

string

string

null

null

x-api-key

Any of:

string

string

null

null

api-key

Any of:

string

string

null

null

Request Body^required

Media typeapplication/json

object

Example ^generated

{}

Responses

200

Successful Response

Media typeapplication/json

Example ^generated

example

422

Validation Error

Media typeapplication/json

HTTPValidationError

object

detail

Detail

Array<object>

ValidationError

object

loc

required

Location

Array

msg

required

Message

string

type

required

Error Type

string

Example ^generated

{
  "detail": [
    {
      "loc": [
        "example"
      ],
      "msg": "example",
      "type": "example"
    }
  ]
}