Guides
How to make an AI video from a text prompt
By the Ofrai team
Start with one short shot. Write the subject, the action and how the camera should move, then check the duration and credit estimate in Ofrai Video Studio before generating.
From prompt to clip
Open Video Studio and choose a model that supports text input. Try: “A slow dolly toward a blue ceramic cup as steam rises in soft window light, one continuous shot.” Pick a supported duration, check the credit estimate and generate. Review the whole clip before downloading it from your library.
Why use one creative studio
Video models improve quickly and differ in style, length, and cost. Ofrai lets you compare options in one place and pay per generation in credits, instead of locking into a single lab’s queue and plan.
Tips for better text-to-video
Describe shot composition and motion explicitly. Start with shorter clips while iterating. If a model supports image references, use them for character or product consistency after a strong still is ready.
Text-to-video versus image-to-video
Start from text when the shot can be described in words, such as a landscape or abstract motion. If you already have an approved product photo, character design or composition, use it as the first frame with a model that accepts image input. Check the finished clip for changes to faces, labels or framing.
Photo-to-video guide: prompts and troubleshootingChoosing a video model by shot length
Model choice matters more than prompt cleverness for most results. Short clips under five seconds are usually the cheapest way to test an idea, and a bad take is inexpensive to discard. Longer clips where a subject has to move convincingly — hands, hair, fabric, liquids — reward models that hold subject identity across a take. Cinematic plates where atmosphere carries the shot reward models with strong lens and light behaviour. Read each model page for its supported duration and credit rate before you commit to a long generation.
Control the credit cost of a video project
Work backwards from your shot list rather than a monthly guess. Assume a rough cut of twenty shots where three shots carry the story. Run exploration on a cheaper model first, then spend the expensive models only on the finals. If your premium model would be more than about a third of projected spend, drafts are on the wrong lane — move exploration down and keep the premium budget for the shots that must be right.
A five-second product shot from a text prompt
A matte black wireless headphone rotates slowly on a white turntable, soft studio light, the camera holds still, clean product commercial.
The prompt gives the product one movement and keeps the camera still. If a clip adds unwanted camera movement, keep the duration and resolution fixed and make the camera instruction more explicit before retrying. See the recorded Kling and Seedance runs for the same prompt.
FAQ
- Can I use the same credits for video and chat?
- Yes. Ofrai uses one credit balance across chat, image, and video.
- Where do finished videos go?
- Completed generations appear in your Ofrai library, where you can download or reopen the studio.
- How long can a generated clip be?
- It depends on the selected model. Each model page lists its supported duration and the credit rate, so check before generating a long clip.
- Can I use an image as the starting point for a video?
- Yes, when the model supports image input. Generate or pick the still first in Image studio, then animate the approved frame for more control over appearance.
More guides
- Photo-to-video guide: prompts and troubleshooting
- One subscription, many models
- DeepSeek online
- Claude online
- GPT-6 Astra pricing, API cost, and token limits
- Best AI video models for text-to-video and image-to-video in 2026
- How to choose an AI model
- How to choose a chat model
- One credit balance for GPT, Claude, Flux & Seedance
- Poe vs OpenArt vs Leonardo — and when Ofrai belongs on the list
- Krater alternatives: all-in-one AI workspaces with shared credits