AI short video generator: a 5 to 10 second clip from a sentence
A few seconds of motion makes a post, ad or web page stand out, but filming it takes time and gear. This prompt has your todo.is agent turn a scene description, or one of your photos, into short MP4 clips you can drop into Reels, Shorts or a video editor.
The prompt
- Make 2 versions of a [LENGTH]-second video clip showing [WHAT HAPPENS IN THE SCENE]. Visual style: [STYLE]. Camera: [CAMERA MOVEMENT]. Format: [ASPECT RATIO], MP4. No text or logos in the video, and keep the main subject in the centre so it can be cropped. If I attached a photo, animate that photo instead and keep it looking like the original. Tell me the exact prompt you used for each version.
What to change
- [LENGTH]: 5 or 10. Clips are a few seconds long; make several for a longer video.
- [WHAT HAPPENS IN THE SCENE]: One clear action, e.g. "steam rising from a fresh cup of coffee on a wooden table at sunrise".
- [STYLE]: E.g. "realistic, warm natural light", "3D animation", "cinematic, moody".
- [CAMERA MOVEMENT]: E.g. "slow push in", "still camera", "slow orbit around the product".
- [ASPECT RATIO]: 9:16 for Reels, Shorts and TikTok; 16:9 for YouTube and websites; 1:1 for feeds.
Example result
- Clips: coffee at sunrise
- 5 seconds · realistic, warm light · slow push in · 9:16 vertical
- Files delivered
- • coffee_steam_v1.mp4: close-up, steam curling up as morning light moves across the table
- • coffee_steam_v2.mp4: wider shot, a hand sets the cup down, then the camera moves closer
- Prompt used for version 1
- "Vertical 9:16 realistic video, 5 seconds. A ceramic cup of black coffee on a worn oak table by a window at sunrise. Soft golden light, thin steam rising and curling. Slow, steady camera push-in toward the cup. Shallow depth of field, warm tones, calm mood. No text, no logos."
- Prompt used for version 2
- "Vertical 9:16 realistic video, 5 seconds. A hand places a ceramic coffee cup on an oak table near a sunny window, steam rising. Camera starts wide and slowly moves in. Warm morning light, natural colours. No text, no logos."
- How to use the clips
- • Reels or Shorts: loop the clip, add your text and music in the app
- • Website header: use the 16:9 version, muted and looping
- • Longer video: make 3 to 6 clips with the same style words and join them in a video editor
- • Voice-over: ask your agent for an MP3 voice-over and add both to your editor
- What to expect
- • Short, simple actions work best; complex scenes with many people may look odd
- • Hands, faces and text in the scene can come out distorted, so check each clip
- • If a version is close but not right, reply with one change, e.g. "less steam, slower camera"
How to do it with todo.is
- Copy the prompt and describe one simple scene, or attach a photo you want to bring to life.
- Paste it on the Today screen in todo.is or send it to your agent with the photo.
- Your agent generates two short MP4 clips and shows you the prompts it used.
- Download the one you like, or reply with a single change to get a new version.
Tips
- Attach a product photo and ask for "a slow orbit" for an easy product ad clip.
- Describe one action per clip. "A wave rolls in" works better than a whole story.
- Reuse the exact prompt your agent shows you, changing one detail, to make a matching series.
- Add text, music and captions in your editing app rather than asking for them inside the clip.
FAQ
- How long can the videos be? Each clip is a few seconds, usually 5 to 10. For a longer video, make several clips and join them in a video editor.
- Does the clip have sound? Some video models can add a soundtrack: ask for “with sound” in your to-do. Otherwise add music in your editing app, or ask your agent for a voice-over MP3 to place on top.
- Can I use the clips in ads? Often yes, but it depends on the video model’s terms, so check them for commercial use. Don’t use real people’s faces or brands you don’t have rights to.
- How many credits does a video use? Videos use more AI credits than images. Your plan includes credits, and you can ask for one version instead of two to use fewer.
JavaScript is required to use the todo.is app.