generated: '2026-07-19'
method: derived
source: openapi/higgsfield-openapi-original.json + https://docs.higgsfield.ai/docs/how-to/introduction
description: >-
  Entity graph derived from the Higgsfield OpenAPI document and the documented
  request/response model. The API is a thin async job surface: a generation
  Request is created against a Model and yields image or video Result assets.
entities:
- name: Model
  description: >-
    A generative model addressed by its model_id path (e.g.
    higgsfield-ai/soul/standard, veo3.1, kling-video/v2.1/master/text-to-video).
    100+ models across image and video families.
  id_field: model_id
  key_families: [higgsfield-ai/*, veo3.1/*, sora-2/*, kling-video/*, bytedance/seedance/*, minimax/hailuo-*, flux-pro/*, reve/*, nano-banana, wan-25-preview/*]
- name: Request
  description: An async generation job created by POSTing to a model endpoint.
  id_field: request_id
  id_format: uuid
  fields: [status, request_id, status_url, cancel_url]
  statuses: [queued, in_progress, completed, failed, nsfw, canceled]
- name: Result
  description: The generated media returned on a completed Request.
  variants:
  - images[]  # each { url }
  - video     # { url }
relationships:
- from: Request
  to: Model
  type: belongs_to
  via: model_id (path)
- from: Request
  to: Result
  type: has_one
  via: images[] / video (on completed status)