Two-level video generation

Video workflow

Pick the approach, then — only on-platform — the method. One clear two-level choice that flows planner → studio, without disturbing the external Veo/Omni flow that already shipped. Light app theme · clickable. No caching — refresh for latest.

Video post Level 1 — approach (planner · exists today)
Google Veo 3.1 external · prompt + paste-back · UNCHANGED
Gemini Omni external · prompt + paste-back · UNCHANGED
On-platform · Higgsfield Level 2 — method (planner · NEW, on-platform only)
Frames-to-video start + optional end frame · frames-capable Higgsfield models
Ingredients-to-video multiple references · Seedance 2.0 + ingredients-capable models
Studio setup gate → clips pre-filled from the post · confirm / override · per-clip inherits
What's new
  • On-platform reveals a frames vs ingredients sub-step in the planner (+ ✨ suggest)
  • The chosen method lands on the studio production (closes today's drop)
  • Studio opens with a pre-filled setup gate; models curated to the method; per-clip config collapses to inherit
What stays untouched
  • The external Veo / Omni prompt + paste-back flow (shipped 2026-06-09)
  • The studio's Higgsfield generation + takes
  • The three productionMethod values; only on-platform gains the sub-step
The pages
Why two levels. "Where/how" (Veo · Omni · on-platform) is a different question from "which on-platform method" (frames vs ingredients) — and the second only matters on-platform. Veo is the external frames option; on-platform frames uses frames-capable Higgsfield models (not Veo). One schema change (a new video.inPlatformMethod field); everything else reuses existing seams. After your sign-off: detailed spec → code.