Text to video: turn a script or idea into a finished short MP4
You have the words (a product idea, a tip, a short story) but no camera, footage or editing skills. This prompt has your todo.is agent split your text into scenes, generate a short AI clip for each one, add a voice-over and subtitles, and join everything into one MP4 you can download.
The prompt
- Turn this text into a [LENGTH]-second video: [YOUR TEXT OR IDEA]. First split it into 4 to 6 scenes and show me the scene list with the line spoken in each. Then make one short AI clip per scene in a [VISUAL STYLE] look, record a voice-over in [LANGUAGE AND VOICE], add subtitles burned into the video, and join it all into one [ASPECT RATIO] MP4. Keep the same main character or product looking the same in every scene. Also give me the subtitles as an SRT file and the voice-over as an MP3.
What to change
- [LENGTH]: 20, 30 or 60. Each AI clip is a few seconds, so 30 seconds is about 5 or 6 scenes.
- [YOUR TEXT OR IDEA]: Paste a script, a blog paragraph, or a one-line idea like "why you should water plants in the morning".
- [VISUAL STYLE]: E.g. "realistic, natural light", "3D cartoon", "watercolour illustration", "cinematic, moody".
- [LANGUAGE AND VOICE]: E.g. "English, calm female voice", "Spanish, upbeat male voice".
- [ASPECT RATIO]: 9:16 for Reels, Shorts and TikTok; 16:9 for YouTube; 1:1 for feeds.
Example result
- Video plan: "Water your plants in the morning"
- 30 seconds · realistic, soft morning light · English, calm female voice · 9:16
- Scenes
- • Scene 1 (0-5 s): close-up of dew on tomato leaves at sunrise. Voice: "Most people water their garden at the wrong time."
- • Scene 2 (5-11 s): a watering can pouring at the base of a plant. Voice: "Morning watering soaks the roots before the heat arrives."
- • Scene 3 (11-17 s): midday sun, a puddle drying fast on hot soil. Voice: "Water in the afternoon, and much of it evaporates."
- • Scene 4 (17-24 s): leaves at night, slightly wet and dark. Voice: "Water at night, and wet leaves invite mould."
- • Scene 5 (24-30 s): a full, healthy vegetable bed, camera slowly rising. Voice: "So set your alarm. Your plants will thank you."
- Files delivered
- • morning_watering_30s.mp4: the finished video, 1080x1920, subtitles burned in at the lower third
- • morning_watering_30s.srt: subtitle file, timed to the voice
- • morning_watering_voice.mp3: the voice-over on its own
- How it was put together
- • Each scene was generated as its own short AI clip from a detailed scene prompt
- • The voice-over was made with text to speech, then each clip was trimmed to match its line
- • ffmpeg joined the clips with short cross-fades and burned in the subtitles in a clean white font with a dark outline
- Before you post
- • Watch it once with the sound off: the subtitles should tell the whole story
- • Check hands, faces and any text in the AI scenes; ask to regenerate a scene if something looks wrong
- • Add music in your editing app or the platform's sound library, at a low volume under the voice
How to do it with todo.is
- Copy the prompt and add your text, length, style, voice and format.
- Paste it on the Today screen in todo.is, or send it to your agent on WhatsApp or Telegram.
- Approve or edit the scene list your agent shows you before it generates the clips.
- Download the MP4, SRT and MP3. Reply with changes like "redo scene 3, less dark" and only that scene is remade.
Tips for a better result
- Write one idea per sentence. Each sentence becomes a scene, so short sentences make cleaner cuts.
- Describe your main character or product the same way in every scene so it stays consistent between clips.
- Keep exact text (prices, names, URLs) out of the AI scenes and ask for it as an overlay; your agent adds it afterwards so it is spelled right.
- Attach a photo of your product or a reference image to anchor the look of every scene.
- Start with 20 or 30 seconds. Short videos cost fewer credits and usually hold attention better on social feeds.
text to video: FAQ
- How long can a text to video be? Each AI clip is only a few seconds, so longer videos are built from several scenes joined together. 20 to 60 seconds works well; longer videos take more scenes and more credits.
- Can I use my own voice instead of an AI voice? Yes. Attach a voice recording and your agent times the scenes and subtitles to it instead of making a text to speech voice-over.
- Will the same person look identical in every scene? Not always. AI video can change small details between clips. Attaching a reference photo and using the same description in each scene helps a lot.
- Does making videos use credits? Yes. AI video clips and voice use the credits included in your plan, and each scene is a separate clip.
JavaScript is required to use the todo.is app.