Summer Sale — 15s Reel

Building clips Video

Each clip is now the deep host of the shared media engine — its own composed-prompt panel & per-clip resource checklist, hosted inside the studio's pink chrome. Build the clip, then generate. One ≤15s Seedance generation per clip.

Saved
Brief & Script Configuration Clip planning 4 Per-clip work 5 Export
Clips (3) 35s
1
Hook · 2 shots
Generated · take chosen
2
Walk & product · 3 shots
Building…
3
CTA packshot · 2 shots
Not started
1 / 3 clips generated 33%
Each clip is its own engine host — pick a model, compose its prompt, fill its resources, then generate (each ≤15s).
Clip 2 of 3 · Walk & product Building
Frames Ingredients
Clip 1 · Hook Per-clip work · clip 2 of 3 Clip 3 · CTA
Clip script — narration · captions · visual direction
Shots — Seedance renders Shot 1/2/3 as ONE coherent ≤15s sequence
Shot 1
Shot 2
2 shots · ~12s · keep within the clip's 15s budget
Composed prompt New · engine
[NEW] The shared media engine, hosted per-clip: a Higgsfield-only ProviderModelPicker, an editable composed VIDEO prompt + guide, a refine row with revertible history, and a clip-scope badge. The composed/refined prompt pins to this clip's snapshot so Generate-clip below runs exactly it. Until you pick a model for the clip, the panel stays disabled.
Shared media engine · clip-scoped
Media provider
higgsfield
On-platform per-clip video generation is Higgsfield-only — google-flow stays in the manual / DIY path.
clip 2 · resolveMediaProviderKey
Higgsfield model — pick one to unlock this clip's prompt
Seedance 2.0 — Multi-shot
Composed video prompt — clip 2
Pick a model first
This clip has no committed model yet, so the composed prompt is disabled. Choose Seedance 2.0 above to compose and refine this clip's VIDEO prompt. Each clip carries its own picker + snapshot.
Preview — what the engine will compose for this clip
15s vertical reel · Aanya walks a sunlit market in the linen shirt, shirt moving naturally in the breeze; medium tracking → close-up on fabric texture; warm naturalistic grade, shallow DoF, product readable; on-screen “Effortless / Breathable linen”…

Compose & refine use the free text path (your AI key / platform key) — never a media credit. Generate clip below spends a media credit and resolves your Higgsfield connector via resolveMediaProviderKey — the composed prompt is pinned to this clip's snapshot.

Preview the unlocked engine (after a model is picked) clip snapshot
Composed video prompt — clip 2 (editable)
Prompt guide
  1. 1.Describe this clip as one coherent ≤15s sequence — its shots, motion, light, grade and the exact product, as if briefing a DP.
  2. 2.Reference the clip's resources below (persona / product / VO) so the model uses the real assets, not look-alikes.
  3. 3.Call out on-screen copy and pacing; keep within the clip's 15s budget.
Refine the prompt
1warmer morning grade
2tighter on the fabric texture
Prompt composed by your AI provider; the clip is delivered by your Higgsfield connector. Pinned to clip 2's snapshot.
Model & settings — this clip inherits the production default; the picker above is the engine's copy
Seedance 2.0 — default
720p
9:16
12s
Reference resources New · engine
[NEW] Replaces the old read-only "Reference ingredients" block with the actionable ResourceChecklist. In Ingredients mode the clip needs persona / product / VO; Frames mode would swap in start frame / end frame roles. Each row = one resource node — role · status · a per-row ⚡ on-platform / ✦ DIY toggle · actions, with an inline ResourceGenerator + candidate grid. A generated resource fills its node; never a nested checklist.
Clip 2 resources · Ingredients mode 1 have · 1 needed · 1 generating
Persona — Aanya identity image
via Pick — aanya-identity.png · keeps face/hair consistent across shots
Have
Attached as a Seedance reference — the model uses the real identity, not a look-alike.
Start frame — hero product in scene image
on-platform generation · seeds Seedance's first frame for this clip
Generating
On-platform DIY
Generate this start frame on-platform, then “Use this” to fill the node. One asset for one node — flat, no nested checklist.
Media provider
higgsfield
resolveMediaProviderKey
Higgsfield model
Nano Banana Pro
Resource prompt — start frame
Prompt guide
  1. 1.Describe the opening frame this clip should start on — composition, light, framing — as a single still.
  2. 2.Attach reference chips (persona / product) so wardrobe and identity stay on-brand; the result fills this node directly.
Prompt composed by your AI provider; the frame is delivered by your Higgsfield connector.
Refine the prompt
1warmer morning light
References for this resource
Persona Product Pick from library Upload
Candidates (3)
“Use this” fills the start frame node and flips it to Have.
Voiceover — Aanya VO audio
prompt / DIY · BYOK voice (ElevenLabs) — synced as a Seedance audio reference
Needed
On-platform DIY
DIY mode hands you the composed VO script + guide with no Generate button — the arm for BYOK voice (ElevenLabs). Bring the audio back via Upload or Pick from library.

Per-resource Generate resolves credentials via resolveMediaProviderKey (your Higgsfield connector) and spends a media credit. Only compose / refine uses the platform text key — that path is free.

Audio
Voiceover · Brand Voice (ElevenLabs)
From the clip script · synced as Seedance audio reference once the VO resource is filled
Pending VO resource Edit
Unchanged Generate-clip + takes are exactly the Stage-4 controls — they run the prompt pinned to this clip's snapshot.
Valid for Seedance 2.0 · 2 shots · 1/3 resources ready · 9:16 · 12s
Seedance 2.0 12s
Take 1 ready
720p · 9:16 · 2 shots
Seedance 2.0 12s
Take 2 ready
720p · 9:16 · 2 shots
Clip 1 is done · finish clip 2's resources, then generate.
Next clip → CTA packshot