How to make an AI video from text
Turning a sentence into a short video used to take a studio. Now it takes a prompt and a couple of minutes. This guide walks you through your first AI video on Dora — no experience needed.
Step 1 — Pick a model
Different models are tuned for different things. Good starting points:
- Gemini Omni — flexible and fast; great default.
- Veo 3.1 — highest-fidelity text-to-video.
- SeeDance 2 — cinematic look.
Not sure? Read Gemini Omni vs Veo 3.1 vs SeeDance 2.
Step 2 — Write a clear prompt
A good text-to-video prompt names four things: subject, action, setting, and style.
"A red fox trotting through a snowy pine forest at dawn, soft mist, cinematic, shallow depth of field."
Tips:
- Lead with the subject and the motion.
- Add camera language ("slow push-in", "aerial", "tracking shot").
- Add lighting and mood ("golden hour", "moody, low-key").
- Keep it to one clear scene per clip.
For 30 ready-to-use examples, see the AI video prompt guide.
Step 3 — Choose your settings
- Aspect ratio: 9:16 for Reels/TikTok, 16:9 for YouTube.
- Duration: start short (4–6s).
- Resolution: 720p for drafts, 1080p/4K for finals.
Step 4 — Generate and refine
Generate, watch it back, and iterate. The fastest way to improve is to change one thing at a time — the camera move, the lighting, or the pacing — rather than rewriting the whole prompt.
Step 5 — Download and share
When you're happy, download and post. Vertical 9:16 clips are ready for Reels, Shorts, and TikTok.
Common beginner mistakes
- Overstuffed prompts. Five competing ideas produce mush. One scene, one mood.
- Wrong aspect ratio for the platform.
- Too long. Short clips look cleaner and cost fewer credits.
Ready to try it? Open Dora and make your first AI video →