generated: '2026-08-27'
method: searched
source: >-
  https://docs.segmind.com/docs/get-started/playground,
  https://docs.segmind.com/docs/get-started/models,
  https://docs.segmind.com/docs/pixelflow/the-editor,
  https://docs.segmind.com/docs/serverless-api/random-seed,
  https://docs.segmind.com/docs/platform/pricing-and-billing
type: playground
test_mode: false
summary: >-
  Segmind ships a browser playground and a visual workflow editor, but no test mode. There
  is no sandbox key, no test-vs-live key prefix, no free simulated response and no fixture
  or trigger tooling — every call in the playground runs a real model and spends real
  credits. This is the standard shape for metered inference: the "sandbox" is a console
  you can experiment in, priced identically to production.
surfaces:
  - name: Model Playground
    url: https://platform.segmind.com
    docs: https://docs.segmind.com/docs/get-started/playground
    description: >-
      Run any Segmind model in the browser. A compose dock, one-click settings, up to three
      generations at once, a session strip, and the ability to edit the raw request as
      JSON before sending it — which makes it a request builder as well as a demo.
    billed: true
  - name: Per-model API tab
    url: 'https://www.segmind.com/models/{slug}/api'
    description: >-
      Each model page carries an API tab listing every parameter the model accepts with
      example code, organised as Mode (Sync / Async) x Language and defaulting to the
      async examples. This is the parameter reference the gateway itself does not publish
      as a spec.
    billed: false
  - name: PixelFlow editor
    url: https://platform.segmind.com
    docs: https://docs.segmind.com/docs/pixelflow/the-editor
    description: >-
      Visual workflow canvas with model search, an asset gallery, a logs panel, run
      controls and undo/redo. Individual nodes can be run on their own using upstream
      nodes' saved outputs.
    billed: true
test_credentials: none
test_values: none
fixtures: none
time_simulation: none
key_prefixes:
  live: SG_
  test: null
  note: One key format, one mode. There is no test-mode key.
cost_control:
  - name: Pinning and caching
    description: >-
      Pinning a node in PixelFlow caches its output across runs, skipping re-execution and
      saving credits while iterating on downstream steps. The nearest thing Segmind
      publishes to a cheap rehearsal loop.
    source: https://docs.segmind.com/docs/pixelflow/guides/pinning-and-caching
  - name: Seed reproducibility
    description: >-
      Pass a positive integer seed for reproducible output, or -1 for random; v1 returns
      the seed actually used in x-seed-value so a random run can be replayed exactly.
    source: https://docs.segmind.com/docs/serverless-api/random-seed
  - name: Failed requests are free
    description: >-
      Only HTTP 200 is charged; every error rolls the credit reservation back in full, so
      a malformed experiment costs nothing.
    source: https://docs.segmind.com/docs/platform/pricing-and-billing
  - name: Free entry credits
    description: >-
      New accounts start on the Flexible plan with a small free-credit balance, reported
      separately from spendable credits as free-credits by GET /v1/get-user-credits.
    source: https://docs.segmind.com/docs/platform/account-and-billing-apis
  - name: Team spend limits
    description: >-
      A per-member spend limit can be set; hitting it returns 406 with a message naming
      the period and the cap.
    source: https://docs.segmind.com/docs/platform/teams
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com