Skip to content
FIGGSHIELD

Notice Figgshield is no longer accepting new sign-ups.

Storyboard to video

Storyboard your video as images first, then animate the frames

Lock every shot as a still before you spend a credit on motion. The frames become your board, and each one becomes a clip.

Sign-ups closed

Why the still comes before the clip

A video is a sequence of decisions: framing, subject, lighting, order. If you make those decisions in motion, every change means re-rendering a clip. If you make them as stills, you are editing cheap images until the shot is right, then animating only what you approved.

Storyboarding as images also gives you something to look at and rearrange before anything moves. Lay the frames side by side, cut one, reorder two, and read the whole piece as a comic strip. That read-through catches pacing problems that are expensive to fix later.

It maps cleanly onto how Figgshield works. You generate each frame as an image, and each image comes back with an id. That id is what you hand to the video step, so the shot you approved is exactly the shot that gets animated.

Draw the board one frame at a time

Write one short prompt per shot and generate it with Nano Banana 2. Keep the descriptions concrete: who or what is in frame, the angle, the light, the background. A board for a product teaser might be three frames: the box closed on a table, a hand lifting the lid, the product standing in soft light.

Generate at 1K or 2K while you are still deciding. Those cost 8 credits each, so a six-frame board is cheap to rough out and cheaper to change. Save 4K for the frames you know are final.

Ask your agent in plain language and it calls the tool for you. The connector lives at https://api.figgshield.ai/mcp, and each finished image returns a generation id you will reuse in the next step.

Animate each approved frame

Once a frame is right, animate it by passing its id to the video step with start_from_generation. The image becomes the first frame of the clip, so the composition and lighting you approved carry straight into the motion. Nothing is redrawn.

Describe motion, not scene. The still already fixed the scene. "Slow push in, steam rising off the cup, background softening" is a job the model can do. Use Kling 3 for the shot you will publish, or Kling 3 Turbo while you are still testing which motion reads best. A five-second clip is 35 credits.

Video is queued rather than instant, so your agent polls until each clip is ready. Work through the board shot by shot, and you end with a set of clips in the order you planned.

Keep the look consistent across shots

The risk with a frame-by-frame board is drift: the product is a slightly different color in shot three, the room changes shape. Fight it in the still stage, where fixes are cheap.

Reuse the same descriptive language for the recurring elements in every prompt, and name specifics rather than moods. "Matte navy bottle, brushed steel cap, warm kitchen counter" repeated across frames holds together better than "the bottle, nice lighting." If one frame nails the look, use it as a reference image for the next.

A reference image you send is used in memory for that generation and never stored, so working with unreleased product shots is fine.

A three-shot teaser, start to finish

Say you want a teaser for a coffee grinder, five seconds per shot. Board it as three stills: the grinder on a counter at eye level, a top-down of beans going into the hopper, a close-up of ground coffee falling into the cup.

Generate the three frames at 2K, review them together, and swap the one that feels flat. Then animate each: a slow push on the first, a downward tilt on the second, coffee falling in real time on the third. Three clips, cut in that order, give you a fifteen-second teaser.

  • Frame 1: grinder at eye level, warm morning light, slow push in
  • Frame 2: top-down, beans pouring into the hopper, slight camera drift
  • Frame 3: close-up, ground coffee falling into the cup, shallow focus

What a board costs, and where the files go

Add it up before you start. Six 2K frames are 48 credits, and animating each as a five-second clip is 35 credits apiece. One credit balance covers both images and video, and pricing is at cost: Figgshield does not mark up the models, so a credit buys the same compute the model provider charges for.

Every finished image and clip sits on a private link for seven days, then it is deleted. Download the board and the clips as you go and keep them in your own storage.

There is no free tier. The $5 Pro trial runs 7 days and includes 150 credits, enough to board and animate a short piece end to end. Cancel before it renews and you pay nothing more. It works with Claude on desktop and in Claude Code, with Cursor and Windsurf, and with ChatGPT in Developer Mode on paid plans. Check your credit balance any time.