Plan video shots with AI
Break a short story into timed shots. Decide what a viewer should learn from each cut and what must look the same when the camera changes position.
Plan video shots with AI by writing down the change the audience needs to see. For a clip about a folding stool, a wide view can establish its height, then a close-up can show the catch releasing. Two shots have explained the mechanism. Another sweeping view might add cost without telling the viewer anything.
The storytelling skill connects these decisions to the available generation tools. It distinguishes a sequence built from separate stills from one that a model can accept as several shots. You can settle the timing and visual references on paper before authorising a render.
What it does
- Give every shot a purpose and an approximate duration
- Specify what should remain recognisable across cuts
- Choose between a reference still and a first or last frame
- Match scene changes to an existing narration
- Record which generated take belongs to each shot
Install and use
- Use a compatible agent and state the duration, format and output you want. Production runs need authorised fal access and media tools; check pricing and use only references you may upload. Anyone with a media link may be able to open it.
- Ask: “Plan three shots for this desk lamp. Show the hinge in the middle shot and leave room at the end for my closing line. Stop at the shot list.”
- Read the line over the proposed timings. If a cut arrives midway through an important action, revise the plan before approving any generation.
FAQ
Source
Made by fal-ai-community. Original skill: storytelling, published on GitHub under the MIT licence.
View on GitHub