How do I get a very accurate prompt for video generations on sora give me a detailed step by step guide.
#sora AI prompting
vocal pawn
zealous groveBOT
Hi @vocal pawn. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
To craft a highly accurate prompt for video generation with Sora, follow these detailed, step-by-step best practices based on OpenAI’s official guidance:
1. Define Your Core Intent
Start by clearly stating what you want to see. Think of this as briefing a cinematographer: if you leave out details, the model will improvise. Decide if you want a tightly controlled result (more detail) or are open to creative surprises (less detail) Crafting a successful video prompt.
2. Structure Your Prompt
Use a consistent structure to organize your ideas. A recommended template is:
[Prose scene description: characters, costumes, scenery, weather, etc.]
Cinematography:
Camera shot: [framing and angle, e.g. wide establishing shot, eye level]
Mood: [overall tone, e.g. cinematic and tense, playful and suspenseful]
Actions:
- [Action 1: a clear, specific beat or gesture]
- [Action 2: another distinct beat within the clip]
Dialogue:
[Short, natural lines if needed]
3. Be Specific and Concrete
- Visual cues: Describe the style and look early (e.g., “1970s film,” “IMAX-scale,” “16mm black-and-white”).
- Camera and motion: Specify shot type, angle, and movement (e.g., “wide shot, low angle,” “slow dolly left”).
- Lighting and palette: Set the mood with lighting and color (e.g., “warm backlight with soft rim”).
- Action: Break down movement into clear beats (e.g., “Actor takes four steps to the window, pauses, and pulls the curtain in the final second”).
- Dialogue and audio: Place dialogue in a separate block, keep lines short, and label speakers for clarity Visual cues that steer the look, Dialogue and Audio, Control motion and timing.
4. Use Image Input for Reference (Optional)
For even more control, provide an image as a visual anchor. The image will be used as the first frame, and your prompt will define what happens next. Make sure the image matches your target video’s resolution and is in a supported format (JPEG, PNG, WEBP) Use image input for more control.
5. Set API Parameters Explicitly
When using the API, set model, resolution, and duration as parameters, not in the prompt text. For example:
model: sora-2 or sora-2-prosize: 1280x720, 720x1280, etc.seconds: 4, 8, or 12
Crafting a successful video prompt
6. Iterate and Refine
Prompting is iterative. Generate a first version, review the output, and refine your prompt for clarity, specificity, or style. Use Sora’s Remix feature to make controlled tweaks without losing what already works Iterate with the remix functionality.
7. Keep It Manageable
- Limit the number of characters and complexity of motion for better fidelity.
- For longer or more complex scenes, break them into shorter clips and stitch them together later Creating videos with Sora.
8. Example of a Detailed Prompt
A sunlit urban park in early autumn. A young woman in a red coat sits on a wooden bench, reading a book. Golden leaves drift down around her. A small dog waits at her feet.
Cinematography:
Camera shot: medium close-up, eye level, shallow depth of field (sharp on woman, blurred background)
Mood: peaceful, warm, cinematic
Actions:
- Woman turns a page and smiles softly
- Dog looks up and wags its tail
Audio:
Soft breeze, distant children playing, leaves rustling