SwishX API
One reasoning API for MLR-ready video generation. Send a prompt, get back an asset with claims linked and fair balance verified, not a separate compliance pass bolted on afterward.
curl https://platform.swishx.com/v1/videos \
-H "Authorization: Bearer $SWISHX_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "finsen-3.2",
"prompt": "A pharma rep opens with a warm, confident smile."
'}
The Reasoning Model Built for Life Sciences
Glorious, compliant and as desired on the first attempt
Five dedicated co-workers shape the brief before production starts.
Each shot gets routed to the model best suited for its motion, tone, and format.
Multi-model routing
On-label by construction
Off-label language, missing fair balance, and unlinked claims are structurally prevented, not caught in review.
Proprietary reasoning, purpose-built for life sciences.
How the brief gets processed
Brief decoded
Context harnessed
Prompts layered
Claims linked
Model routed
Output verified
Named after the scientists whose work made it possible.
Every two or three versions, SwishX retires a model name and starts a new one, each after a Nobel laureate in Physiology or Medicine. Behring, then Ross, now Finsen. Any finsen-3.x version stays callable indefinitely.
- Finsen 3.2Aug 2026Current
Sharper claim-grounding accuracy and lower multi-model routing latency than 3.1.
- Reduced false-positive rate on the claims-linked stage for comparative claims.
- Faster model-routing decisions when a downstream model is near capacity.
- Finsen 3.1Jun 2026
Keyframe-first generation, reducing hallucination on rare disease states and off-label claims.
- Introduced keyframe_id on video generation, reuse or regenerate a still without re-rendering the clip.
- Meaningfully fewer off-label visual implications on longer (12–16s) generations.
- Finsen 3.0May 2026
Full reasoning pipeline shipped: brief decoded, context harnessed, prompts layered, claims linked, model routed, output verified.
- First version where request/response shape is considered stable going forward.
- Output verification added as a distinct post-render stage.
- Ross 2.5Mar 2026
Claim Extractor and Audience Classifier split into dedicated reasoning stages, instead of being handled implicitly inside context harnessing.
- Claim-linking rejections now include which specific claim triggered them.
- Ross 2.2Feb 2026
Compliance Verifier added: automated ISI and fair-balance checks before render, not just at claim-linking time.
- First version to reject a request for fair-balance reasons pre-generation, not just unlinked claims.
- Ross 2.0Jan 2026
Multi-model routing across Veo, Kling, and Runway; scene-level model selection instead of a single fixed generator.
- routed_model first appears in the response shape.
- Behring 1.5Dec 2025
Introduced the Context Harness: brief, dossier, and regulatory ruleset merged before generation.
- Behring 1.0Nov 2025
Text-to-video orchestration, single-pass prompting, no compliance layer.
- The first version of the API, a single generation call with no claim-linking or verification stages.
Three calls from prompt to review-ready.
No SDK required to start, just plain HTTPS and JSON, with SDKs for Python and Node if you want them.
Create an API key
Generate a scoped key from the API dashboard. Every key is tied to one project and its own rate limits, so a leaked key never sees another team’s data.
# In the API dashboard → API Keys
export SWISHX_API_KEY=sx_live_9f2c...
Send a request
One call takes a prompt and your Brand Dossier ID through the full reasoning pipeline: context harnessed, claims linked, model routed.
POST /v1/videos
{
"model": "finsen-3.2",
"dossier_id": "dsr_4c1a",
"prompt": "Open on the injection-site demo."
}
Get a verified asset back
The response carries the finished asset alongside exactly which claims it used and which model rendered it, not just a raw file.
200 OK
{
"status": "succeeded",
"routed_model": "veo-3",
"claims_linked": 6,
"compliance": { "verified": true }
}