Introduction
What the SwishX API is, and how it's put together.
SwishX is a reasoning model for life-sciences video — not a thin wrapper around a text-to-video model, but an orchestration layer that decodes a creative brief, grounds it against your brand's claims and regulatory ruleset, routes it to the right downstream generation model, and verifies the output before it's ever handed back to you.
SwishX names each model generation after a Nobel laureate in Physiology or Medicine, rotating to a new name every two or three versions. The current generation is Finsen — the current version is Finsen 3.2 — see the changelog for the full version history back to Behring 1.0.
Who this is for
These docs are for engineers integrating the SwishX API directly. If you're looking for the product itself, see swishx.com; to manage your account, keys, and billing, see the SwishX API Platform.
Why a reasoning model, not a video API
Most video-generation APIs take a prompt and return a clip. That's fine for a demo; it falls apart the moment the output has to survive Medical/Legal/Regulatory (MLR) review. SwishX's reasoning model sits in front of the actual generation step and runs every request through six stages before anything renders:
- Brief decoded — your prompt is parsed into an intent: subject, setting, tone, call to action.
- Context harnessed — your brand dossier (claims, prior assets, tone) and regulatory ruleset (ISI, fair balance, label) are merged with the brief.
- Prompts layered — a structured generation prompt is built from the merged context, not the raw brief alone.
- Claims linked — every claim implied by the prompt is checked against your dossier's approved claim set.
- Model routed — the request is sent to whichever downstream model (Veo, Kling, Runway, Sora) suits the shot.
- Output verified — the rendered result is checked again before the request resolves as
succeeded.
Read more in The reasoning pipeline.
Base URL
All API requests are made to:
https://api.swishx.com/v1