What is text-to-video?
Text-to-video is a type of generative AI that reads a written description — a prompt — and produces moving footage that matches it. Modern models don't just animate stills; they render coherent scenes with motion, lighting, and, in Veo 3.1's case, native synchronized sound. You don't need editing software, a camera, or any technical background. If you can describe what you want to see, you can make it.
Step 1 — Write a clear prompt
A good prompt answers five questions: who or what is in the shot, what they're doing, where it happens, how the camera sees it, and the mood. Keep it to one focused paragraph. Compare these:
Too vague
"A dog in a park."
Clear and specific
"A golden retriever bounding across a sunny park, slow-motion tracking shot from the side, warm late-afternoon light, joyful and cinematic."
If you want spoken words, put them in quotes inside the prompt — Veo 3.1 will voice them in sync. Avoid stacking ten ideas into one prompt; one clear scene per clip produces the cleanest result.
Step 2 — Choose a Veo 3.1 model
Open the veo3gen dashboard and pick a model based on your priority:
- Quality —
veo-3.1-generate-001for the most detailed, polished output. - Fast —
veo-3.1-fast-generate-001for quicker, cheaper iteration while you dial in a prompt. - Lite —
veo-3.1-lite-generate-001for the lowest cost when you just need a draft.
Beginners usually start with Fast to experiment, then re-render the winner in Quality. The full lineup, including the Veo 3.0 models, is in the model reference.
Step 3 — Set length and aspect ratio
Choose a clip length of 4, 6, or 8 seconds — shorter clips iterate faster and cost less. Then pick an aspect ratio that matches the destination: 16:9 for YouTube and websites, 9:16 for TikTok, Reels, and Shorts. veo3gen renders up to 1080p on non-Lite 16:9 outputs and 720p otherwise, so select the resolution that fits your platform and budget.
Step 4 — Generate the video
Press generate. veo3gen sends your prompt to the model and renders the clip, including native synchronized audio — no separate sound step required. Generation takes a short while depending on the model and length; Fast and Lite return sooner than Quality. Because billing is pay-as-you-go at 1 credit = $0.01, you can see exactly what each clip costs before you commit to a bigger batch. New here? The quick-start guide walks through your very first render.
Step 5 — Review and download
Watch the preview. If something is off — the framing, the motion, a detail — tweak that one element in your prompt and regenerate; iteration is how you get great results. When you're happy, download the MP4 and use it anywhere. That's the whole loop: describe, generate, refine, download.
Tips and pitfalls
- Do describe camera movement ("slow dolly in," "handheld") for more cinematic results.
- Do iterate one variable at a time so you learn what each change does.
- Don't pile multiple scenes into a single prompt — generate them separately.
- Don't expect a 4-second clip to tell a whole story; chain clips for longer pieces.
- Start on Fast or Lite to save credits, then finalize on Quality.
Frequently asked questions
Do I need any software or editing skills?
No. veo3gen runs in your browser. You write a prompt, choose a few settings, and download the finished clip — no editor or experience required.
Does the AI video include sound?
Yes. Veo 3.1 generates native synchronized audio, so dialogue and ambient sound come with the clip. Put spoken lines in quotes to have them voiced in sync.
How much does one video cost?
veo3gen is pay-as-you-go at 1 credit = $0.01, so a short clip costs only cents. There are no subscriptions to start.
Turn your first sentence into video
Write a prompt, pick a model, and watch Veo 3.1 bring it to life — with sound — in minutes.