How do I turn a script into a shot list with AI?
Paste the script into a text model and ask it to break each scene into numbered shots, each with a framing, one main camera or subject move, the action and a rough duration. Then check the list against the story before you generate any images. On Gulab the script, the shot list and the storyboard sit on one canvas: a Text node holds the script, a Generate Text node writes the shot list, and each shot becomes an Image node you can later animate.
What a good AI shot list contains
A shot list is only useful if each line can be shot, or generated, on its own. Ask the model for these fields per shot:
- Shot number and scene, so shots can be reordered and regenerated by number.
- Framing: wide, medium, close-up, insert, over-the-shoulder.
- One primary move: a single camera move (push in, pan, handheld follow) or a single subject action. Shots that try to do two things at once are the ones image and video models get wrong.
- Action and dialogue: what happens and what is said.
- Duration: a rough length in seconds, which also tells you how many clips the scene needs.
- Characters and location: who and where, so you can attach the right references later.
Step by step on Gulab
- Drop a Text node on the canvas and paste the script, or describe the idea in the canvas assistant and let it build the nodes for you.
- Wire the Text node into a Generate Text node. In its instruction, ask for a numbered shot list with the fields above, one shot per line, and pick the text model you want.
- Read the list as a story. Cut shots that repeat a beat, split shots that carry two moves, and make sure every scene has an opening shot that sets up the space.
- Label each scene with a Heading node and group its shots inside a Zone node, so the board reads like a script breakdown.
- Create one Image node per shot, wire in your character and location references (or pick them from Elements) and paste the shot line as the prompt. That gives you a storyboard panel per shot. The full panel-to-video process is in AI storyboard to video.
Breaking down a reference instead of a script
If you are working from an existing edit, such as a trailer you want to study or a sequence you want to remake, the Scenes node splits a video into scenes, each with a title, prompt, timing, style, camera movement and a captured frame. That is a shot list written from footage, ready to restyle shot by shot.
From shot list to animatic and final cut
Once the panels exist, you can time them into an animatic to check pacing before spending on video (see AI animatic maker), then connect each panel to a Video node and assemble the clips with Concat, Audio Mix and Export.
Pricing
Gulab bills per generation from one credit balance, and every plan includes every model. Current plan and per-model prices are on gulab.ai/pricing, which is always the live source.
FAQ
Can AI write a shot list from a script?
Yes. On Gulab you paste the script into a Text node and wire it into a Generate Text node with an instruction to return numbered shots with framing, one camera or subject move, action and duration. The canvas assistant can also build this for you.
What should each line of a shot list include?
Shot number, scene, framing, one primary camera or subject move, the action and dialogue, a rough duration, and the characters and location in the shot.
Can I turn the shot list into storyboard images?
Yes. Each shot becomes an Image node with your character and location references attached, so the shot list turns directly into storyboard panels on the same canvas.
Can I make a shot list from an existing video?
Yes. The Scenes node splits a video into scenes with a title, prompt, timing, style, camera movement and a captured frame for each.
Which AI models can I use for the shot list?
Gulab includes text models from several providers in every plan, and you choose one in the Generate Text node. Image and video models for the panels and clips are included too, billed per generation.