How to Make Videos with AI in 2026: Step-by-Step Workflow With Real Costs
Choose your settings and review the credit cost before generating.
Create your first videoMaking a video with AI means describing a shot, letting a model render it, and then doing the part the model can't: checking it. This guide walks through the whole workflow in Viralance, with the real settings each model accepts, what each choice costs in credits and dollars, and where AI video is the wrong tool.
What you need
- One idea for one shot. A single subject doing a single thing. Most models render 4–15 seconds; Seedance 2.5 is the only one here that goes to 30 seconds in one take.
- A Viralance account. New accounts get 4 free credits after email verification.
- A budget for attempts, not just one clip. Plan for at least two attempts per shot; results vary between runs of the same prompt.
No camera or editing software is required, but you do need to watch the output closely before publishing.
Step 1: Pick the input that matches your goal
| You have | Use | Why |
|---|---|---|
| Only an idea | Text-to-video | The model invents the whole scene |
| A photo of a real product, place or person | Image-to-video | The photo anchors the look; text alone can't reproduce a real label |
| Several photos of the same subject | Reference-to-video (VEO 3.1 Fast, Seedance 2.0/2.5, Kling 3.0 Pro, MiniMax H3) | Keeps a character or product consistent across a new scene |
| An existing clip | Edit tools | Subtitles, reframing, background removal, lip sync, upscaling |
Step 2: Start from a template, then change it
The create page has starter templates grouped as UGC & Ads, Product, Cinematic, Character, Lifestyle and Effects (for example UGC Product Review, Unboxing, Before / After, Food Hero, Drone Reveal). A template fills in a detailed prompt, a model, a duration and a resolution. Replace the {PLACEHOLDER} parts with your specifics.
Check the cost before you press Generate: templates inherit the price of their model. UGC Product Review uses Seedance 2.0 at 720p for 5 seconds (12 credits), while 30s Story Ad uses Seedance 2.5 at 720p for 30 seconds, which is 142 credits (about $25.70 even at the Business pack rate).
Step 3: Write the prompt
A prompt that a model can follow names six things:
- Subject: who or what is in frame
- Action: one thing that happens
- Camera: static, slow push-in, orbit, handheld
- Lighting: direction and quality
- Style: realistic, commercial, anime, UGC selfie
- Exclusions: "no text, no extra objects, no people"
Example used in our September test (VEO 3.1 Lite, 4s, 720p, 9:16):
A quiet rocky coastline at sunrise, gentle waves washing around one foreground rock. Slow steady camera push toward the ocean, natural golden light, clear horizon, no people, no text.
Two things a prompt can't do: it won't override the output settings (typing "4K, 60fps" doesn't change resolution), and it won't make a model spell long brand text reliably.
Free and paid helpers on the create page:
- Enhance: rewrites your prompt into several more detailed options for you to choose from. Free (0 credits). Read the rewrite; it can add details you didn't intend.
- Hooks: generates opening-line ideas and inserts the one you pick at the top of the prompt. 1 credit per batch.
Step 4: Choose a model with the right limits
Every figure below comes from the live Viralance catalog (September 2026). "Aspect ratios" are the options offered for text-to-video; image-to-video follows the shape of your source image.
| Model | Durations | Resolutions | Aspect ratios | Credits per clip | Good for |
|---|---|---|---|---|---|
| VEO 3.1 Lite | 4, 6, 8s | 720p, 1080p | 9:16, 16:9 | 3–8 | Cheap drafts with audio |
| VEO 3.1 Fast | 4, 6, 8s | 720p, 1080p | 9:16, 16:9 | 6–12 | Clean all-rounder with audio |
| Kling 3.0 Standard | 5, 10, 15s | 1080p | 9:16, 16:9, 1:1 | 8–22 | 1080p motion at a lower price |
| Kling 3.0 Pro | 5, 10, 15s | 1080p | 9:16, 16:9, 1:1 | 10–28 | Cinematic motion, multi-language audio |
| Seedance 2.0 | 5, 10, 15s | 480p, 720p, 1080p | 6 options incl. 1:1, 21:9 | 10–82 | Talking and lip-synced clips |
| Seedance 2.5 | 5, 10, 15, 30s | 480p, 720p | 6 options | 11–142 | Single takes up to 30s |
| Happy Horse 1.1 | 5, 8, 10s | 720p, 1080p | 9:16 | 8–19 | Native audio and lip sync |
| Hailuo 2.3 | 6, 10s (1080p: 6s only) | 720p, 1080p | 9:16 | 3–7 | Physics and anime, low cost |
| Wan 3.0 | 5, 10, 15s | 480p, 720p, 1080p | 5 options | 4–38 | Native audio, up to 15s |
| MiniMax H3 | 5, 10, 15s | 768p, 2K, 4K | 6 options | 3–24 | 4K output with stereo audio |
| PixVerse V6 | 5, 8s | 720p, 1080p | 9:16 | 4–10 | Budget clips with audio |
| Grok Imagine | 5, 6, 10, 15s | 480p, 720p | 5 options | 2–10 | Cheapest quick tests |
| Vidu Q3 Turbo | 5, 8, 10, 16s | 540p, 720p, 1080p | 5 options | 2–12 | Budget, up to 16s |
The full list, including image-only and editing models, is on the models page.
Step 5: Set duration, resolution, aspect ratio and audio
- Aspect ratio: 9:16 for TikTok, Reels and Shorts; 16:9 for YouTube; 1:1 for feeds. Pick it here rather than cropping later.
- Resolution: draft at 720p, then re-run the winner at 1080p if the model offers it. On VEO 3.1 Fast, 8 seconds costs 12 credits at either resolution, so there the trade-off is only a longer wait and a larger file.
- Audio: VEO, Kling 3.0, Kling O3 Pro, Vidu Q3 Turbo and Seedance models have an audio switch. Turn it off if you'll add your own voiceover.
The credit cost updates as you change settings and is shown before you generate.
Step 6: Budget in credits and dollars
Credits come in one-time packs (no subscription, no expiry): Mini $9/30, Plus $19/70, Starter $29/110, Creator $79/330, Pro $149/770, Business $299/1,650. That's $0.30 per credit on Mini down to about $0.18 on Business.
| Clip | Credits | Cost on Mini | Cost on Business |
|---|---|---|---|
| VEO 3.1 Lite, 4s, 720p | 3 | $0.90 | $0.54 |
| Kling 3.0 Standard, 5s | 8 | $2.40 | $1.45 |
| VEO 3.1 Fast, 8s, 720p | 12 | $3.60 | $2.17 |
| Seedance 2.0, 10s, 720p | 24 | $7.20 | $4.35 |
What the 4 free credits cover: one 4-second VEO 3.1 Lite draft at 720p (3 credits), or two 5-second Grok Imagine 480p or Vidu Q3 Turbo 540p tests (2 credits each).
A first purchase made within 24 hours of signing up gets 5–15% bonus credits depending on the pack; after that, every pack gets +10% for returning customers. See pricing.
Step 7: Review the result before you retry
Watch the full clip, then pause at the start, middle and end.
| Check | Pass condition | If it fails |
|---|---|---|
| Subject | Same subject throughout, nothing extra | Add "no additional objects" |
| Motion | The one action you asked for | Simplify to a single action |
| Shape and text | Objects keep their shape; short text stays readable | Reduce camera motion or start from a photo |
| Framing | Nothing important cut off | Change aspect ratio or composition |
| File | Right dimensions, audio on or off as needed | Check the settings, not the prompt |
Change one thing per retry. If you change the prompt, model and duration together, you won't know which change helped. Our September delivery test shows why: the same bottle prompt gave a bottle without a cap on one run and one with a cap on the next.
If a generation fails on our side, its credits are refunded automatically.
Step 8: Caption, finish and post
- Caption generator: writes a short caption (about 150 characters) with 6–9 hashtags from your prompt. Free.
- Auto Subtitles: burns animated word-by-word captions into the video. 3 credits, for clips up to 5 minutes, in 8 languages.
- Reframe: rebuilds the edges to change 9:16 ↔ 16:9 ↔ 1:1 without cropping. Clips up to 10 seconds, from 6 credits.
- Multi-shot: the Film tool storyboards 4–12 shots from one idea and stitches them (per-shot cost plus 5 credits to assemble).
Viralance does not post to social platforms for you. Download the MP4 and upload it in the TikTok, Instagram or YouTube app.
How long does it take?
In our September 2026 test, six VEO 3.1 Lite 4-second 720p jobs took 79–117 seconds from submission to a ready file, including polling delay. Longer, higher-resolution or busier models take longer, and provider queues vary. Check your library before resubmitting a job that is still processing.
When AI video is the wrong choice
- Exact product claims. If the video must show a real label, ingredient list or dimensions accurately, film the product or use image-to-video and check every frame.
- Real people. Don't generate a recognizable person without their consent. Viralance screens uploads for impersonation and asks you to confirm consent for image-to-video.
- Long talking content. Most models top out at 8–15 seconds per clip. For a 2-minute presenter video, a dedicated avatar tool or a real recording will be simpler.
Frequently asked questions
How much does it cost to make an AI video? From 2 credits (a 5-second Grok Imagine 480p or Vidu Q3 Turbo 540p clip) to 142 credits (a 30-second Seedance 2.5 720p take). A typical 4–8 second 720p clip is 3–12 credits, or $0.90–$3.60 at the $9 Mini rate.
Do I need editing skills? No, but you need to review what the model produces. Captions, subtitles and reframing are built in; combining several clips needs the Film tool or your own editor.
Which model should a beginner start with? VEO 3.1 Lite at 4 seconds and 720p (3 credits) for cheap drafts. Move to VEO 3.1 Fast or Kling 3.0 once your prompt works, and to Seedance 2.0 if someone needs to speak on camera.
Can I make a video of my own product? Yes. Use image-to-video with a clear product photo, keep the motion gentle, and check the label in every frame. See the label checklist.
Can I use the videos commercially? Viralance's terms state you retain ownership of what you generate. You're still responsible for the rights to anything you upload and for claims your ad makes.
Keep going
- How do I animate a product photo?
- How do I add a voiceover?
- How does the TikTok algorithm rank videos?
- How do I generate AI images to animate?
Make a first draft. Open a 4-second VEO 3.1 Lite draft: the link selects the model and duration but does not generate anything until you confirm.
