guides

Generate your first video

Brief in, directed sequence out, with the plan visible before any spend.

Last updated 2026-07-10 · beta

Prerequisites

A Cadenza account (sign in)
For the API path: a key with the generate scope (authentication)

Plan first, free

curl

curl -X POST https://cadenza.media/api/v1/story_intelligence_report \
  -H "Authorization: Bearer cadk_..." -H "Content-Type: application/json" \
  -d '{"raw_intent": "15 second premium launch film for a ceramic pour-over kettle"}'

Read the shot plan and the advisory notes. Iterate on the brief until the plan reads right. This costs nothing.

Generate

curl

curl -X POST https://cadenza.media/api/v1/generate_video \
  -H "Authorization: Bearer cadk_..." -H "Content-Type: application/json" \
  -d '{
    "raw_intent": "15 second premium launch film for a ceramic pour-over kettle",
    "director_mode": "auto",
    "shot_count": 4
  }'

Read the response

status: ok, blocked, or render_failed.
mode_selection: which grammar was chosen and why.
shots[]: per shot, the narrative purpose, camera, routed model, attached reference ids, render result, and provenance record.
story_report: the five advisory checks.

Beta: During the beta, renders are previews while live provider rendering is wired. The plan, routing, continuity, and provenance behavior are the production paths.

Errors