Learn how to write prompts and narrations to generate amazing videos.

In this article, we’ll show you how to write an Action Prompt and Narration.

Action Prompt is the prompt or script that tells the AI what should happen on screen. It describes the visual elements of the scene, including:

Think of Action Prompt as of the visual script for your video.

Pro Tip: You can use an AI assistant, such as ChatGPT, Gemini, or another tool, to help you create a prompt that reflects your request and desired result.

For example, in ChatGPT, we described the video we wanted to generate and asked it to create a detailed prompt and narration. Please note that the prompt must not exceed 3,500 characters. You can ask ChatGPT or any other AI tool you use to generate a detailed prompt while ensuring it stays within this limit.

“I have an image of a ballet dancer and a hip-hop dancer dancing in a garden among the trees. The ballet dancer is holding a bag, which is my product.

I want to create a short marketing video to show that the bag is perfect for any situation—even dancing in a tutu next to a hip-hop dancer in a garden.

The ballet dancer should perform ballet, and the hip-hop dancer should perform hip-hop. Their movements should match and flow well together.

The bag should stay clearly visible throughout the video. Its original design, shape, color, and details must remain exactly the same as in the original image. Please create a prompt for generating a 15-second video on an AI video generation platform. Make sure that the prompt is detailed but not longer than 3,500 characters. Also, please write creative text that matches this video.”

Here is the script we received from ChatGPT:

“Create a premium 15-second fashion advertisement using the uploaded image as the exact reference. A ballet dancer in a tutu holding the featured bag dances alongside a hip-hop dancer in a beautiful, sunlit garden surrounded by trees. The mood is joyful, stylish, unexpected, and effortlessly confident.

CRITICAL PRODUCT PRESERVATION: The bag must remain clearly visible and identical to the original image in every frame. Preserve its exact design, shape, proportions, color, material, texture, handles, straps, stitching, hardware, logo, and all details. Never redesign, distort, recolor, replace, duplicate, or deform the bag. Keep it naturally held, facing the camera as much as possible, unobstructed by hands, clothing, or body movements. Product accuracy takes priority over complex choreography.

CHOREOGRAPHY AND TIMING:

0–3 seconds: Establish the lush garden in warm golden-hour sunlight. The ballet dancer begins graceful arm extensions, pointed-toe steps, and an elegant turn while holding the bag visibly. The hip-hop dancer starts with relaxed, rhythmic grooves.

3–7 seconds: The dancers perform complementary choreography to the same imaginary beat. The ballet dancer executes controlled pivots, delicate footwork, flowing arms, and light relevés. The hip-hop dancer responds with smooth body waves, shoulder pops, rhythmic footwork, and confident isolations. Their movements feel coordinated and connected while retaining their distinct styles.

7–11 seconds: Build the energy through a playful dance exchange. Use mirrored gestures, synchronized accents, and fluid transitions. The ballet dancer glides and pivots while keeping the bag stable, naturally positioned, and visible. The hip-hop dancer adds dynamic but controlled street-dance movements. The tutu flows naturally without covering the bag.

11–15 seconds: Both dancers finish in confident, complementary poses. The ballet dancer subtly presents the bag toward the camera, with its front and details clearly visible, while the hip-hop dancer strikes a stylish pose beside them. End with a polished product-focused composition, the bag sharp and prominent against a softly blurred garden.

CAMERA AND STYLE: Cinematic fashion campaign combining classical elegance with modern streetwear. Realistic anatomy, natural fabric movement, warm sunlight, lush greenery, shallow depth of field, smooth stabilized tracking, and a gentle push-in toward the final product shot. Keep both dancers and the entire bag in frame. Maintain consistent faces, clothing, body proportions, and bag appearance throughout. Movements must be fluid, physically plausible, and continuous.

AVOID: Bag alterations, obscured product details, cropped bag, morphing, flickering, jitter, distorted hands, extra limbs, abrupt transitions, excessive camera shake, and unrealistic physics. Do not generate text or additional logos.”

Since this prompt meets our needs, we copy the text and paste it into the Action Prompt field. You can modify the generated prompt to better suit your specific requirements.

You can manually edit your Action Prompt or automatically adjust it to match the newly selected image.

Once you have written or pasted your prompt, you can easily edit it in the prompt editor. Simply click the small eye icon next to the Action Prompt title to open a pop-up window, where you can edit your text and search for specific words.

If you decide to change your prompt while editing your settings, you don’t need to do it manually. You can adjust it automatically. Simply click the Adjust Prompt to Image button, wait a few seconds, and check the updated prompt.

Narration is the dialogue or narration that the character speaks in the video. Whatever you enter in this field will be spoken by the character or narrator.

You can add several narrators/speakers if needed.

Note. Please note that not all AI models support voice-overs or music generation. For example, Runway Gen 4 Turbo and Runway Gen 4.5 do not offer audio generation. Available features depend on the AI model you select.

Note. Our license does not cover videos in which children are speaking. If you add a voice to a child/children, the resulting video will not be covered by our license. Please check this article for more information.

In the “Speech / Narration” section click on the narrator drop-down list and add new narrator by clicking “Create speaker”. Click green check mark to save the change.

Note. Once you create a new narrator, you cannot delete it from the list of this particular video generation. However, you can remove it from the list of active narrators so that it will no longer speak. It will still remain in the list.

By hovering over the existing narrator, you will see 2 options: to rename it or add narrator to the video.

To rename the Narrator, hover over it, click pencil icon, change the name, and click green check mark to save the change.

If you want to add additional narrators to the video, simply click “Add speech box” and additional “Speech” field will appear. You can add text to each of those fields and selected AI models will vocalize them.

You can also easily remove Speech box by clicking Trash icon from the right side of the selected Speech box.

If you don’t have time or ideas for writing narration, you can use an AI assistant, such as ChatGPT, Gemini, or another tool, to help you create it. :)

Simply click the Speech/Narration text field and type or paste the text you want your narrator(s) to say.

We have added 3 Narrators:

Narrator — “Grace has no limits. Rhythm has no rules. From pirouettes to street moves, take your style everywhere.”

Woman in White Ballet Dress — ”One bag. Endless possibilities.”

Man in Orange Beanie — "Carry your world. Own every moment.”

Before moving on to the next tutorial articles, take a moment to preview the final video we created using these prompts and narrations.

Already watched the preview? Great! Let’s move on to the next tutorial to learn how to Select AI Model and Output Parameters.

If you have any questions, feel free to contact us via in-app chat or email at [email protected].

We look forward to seeing your brand thrive with ActiveCast videos!