# Prompt a shot from your own first frame

Recipes / Prompt guide / Which model. Checked 2026-10-13.

Prompts that keep the camera, the picture and the acting apart, so you can swap one without breaking the others.

- Use when: Every image-to-video job and every still built shot by shot.
- Phrase: start frame attached; then four separate parts: the subject, the place, the camera move, the look; and the action of the person on its own line
- Page: https://directorchair.ai/guide/guide-shot-from-first-frame
- JSON: https://directorchair.ai/api/guide/guide-shot-from-first-frame.json

## Watch out

- In image-to-video, spend the words on what happens; the start frame already carries the look, the identity and the composition. (measured; Director Chair field notes on video prompting)

## The parts, kept apart

1. **First frame or blocking.** Attach your start image, or block the frame first in a 3D stage (Camera Control) and export it as the composition reference.
2. **Subject** and **place**: only what the image does not already show.
3. **Camera move:** one move, written as a move card (camera, pace, keep, ends).
4. **Action:** what the person does, as behaviour with a trigger.
5. **Look:** the job's look sentence, the same in every prompt.
6. **Sheets:** say whether a character sheet or an environment sheet is attached, and give each its role.

Both free tools we point to work this way: the move library keeps the camera line apart from the scene, and the 3D stage asks for subject, environment, camera move and look as separate fields.

## Prompts

### Any chat model

```
Give me a filled-in prompt that follows this guide, plus the one thing to check in the result, for a short video about [your product], for [your audience]. The technique is prompt a shot from your own first frame: prompts that keep the camera, the picture and the acting apart, so you can swap one without breaking the others. How it works: start frame attached; then four separate parts: the subject, the place, the camera move, the look; and the action of the person on its own line. It is right for: every image-to-video job and every still built shot by shot. In image-to-video, spend the words on what happens; the start frame already carries the look, the identity and the composition. Be concrete: numbers, seconds and settings, no theory. Ask me one question first if something you need is missing.
```

## Look further

- https://aicameramovements.com/ : a looping example clip of this move and a free prompt card (link only)
- https://aicameracontrol.com/ : a free 3D stage to place the person and set the lens before you generate

## Goes well with

- Framing label: https://directorchair.ai/composition/framing-label
- What moves and what stays: https://directorchair.ai/motion/what-moves-what-stays
- The look block: https://directorchair.ai/look/global-look-block

## Source

Structure as on the aicameramovements prompt builder and the Camera Control fields (both read 13 Oct 2026), in our own words

Text CC BY 4.0, Director Chair (directorchair.ai). Reference clips and their card texts belong to their makers.