AI Video Generator

Only partly. Generating needs an account with a confirmed e-mail address, so there is no way to make a clip without signing up. You get 10 free credits at the start with no card, but the cheapest clip costs 30 credits (LTX-2, 6 s), so your first video means buying a pack. The “Enhance prompt” button is free, and a failed generation costs nothing. Credits never expire, and a monthly subscription is optional.

Type the scene in the Prompt box in 1–3 sentences: who, where, what they do and how the camera moves, for example “A woman in a yellow coat walks down a wet neon street at night, the camera slowly follows her, rain and light reflections in the puddles”. Then pick a model, length and aspect ratio and click “Generate video”. Every one of the 16 models works from text alone. If you only have a rough idea, “Enhance prompt” expands it for free. For ideas for whole videos without a face, see the guide to faceless YouTube videos.

Upload a photo in the “Starting image” field (JPG, PNG or WebP up to 10 MB) and describe what should move, for example “the person blinks and smiles slightly, camera still, face unchanged”. 13 models accept a photo. Veo 3.1, Veo 3.1 Lite and Wan 2.7 are text only, and if you pick one of them the generator shows a message and takes no credits. Kling 3.0 Standard (50 credits for 5 s) is a good choice for portraits, and Seedance 2.5 can take up to 50 reference images at once. More about animating photos is on the AI photo animation page.

Yes. You can write the scene in your own language: we translate it into English before it goes to the model, and text in quotation marks, for example a line to be spoken, stays unchanged. The Language menu offers English, Polish, German and Spanish, and only Polish changes anything: it narrows the model list to the 4 models with good Polish speech (Gemini Omni Flash 1.1 and three Veo versions). For the other languages the list stays full, so check the first result before you order longer clips. It is a filter on the list, not a separate generation setting.

The generator quotes 1–3 minutes per clip, and the biggest models, such as Seedance 2.5, can be slower. Credits are taken at the start. If a generation fails, the content check blocks the result, or the job does not finish within 30 minutes, your credits come back automatically. The generator asks you not to close the page, but if you do, the server checks unfinished jobs every 10 minutes and saves the finished clip in your account history.

Gemini Omni Flash 1.1 and the three Veo versions make a fixed 8 seconds, Wan 2.7 and Grok Imagine make 5 seconds, the other models usually 5–10 s, and Seedance 2.5 goes up to 30 s. Gemini Omni Flash 1.1, Veo 3.1, Veo 3.1 Fast and Seedance 2.5 generate sound together with the picture. LTX-2, LTX-2.5 and Veo 3.1 Lite make clips without sound, and the descriptions in the generator also mention audio for Kling 3.0 Pro and optional audio for PixVerse v6. More about models with sound is in the guide to an AI video generator with sound, and a tool comparison is in the video generator ranking.

In general yes, within what the Terms and the AI model providers’ terms allow, for example in a post, an ad or on a shop page. The service allows safe content only (SFW): your prompt and uploaded photo are checked before generation and the finished clip is scanned, so a blocked generation ends with a refund. Don’t upload photos of people who haven’t agreed to it. The details are in the Terms.

Here you have full control: you pick the model, length and aspect ratio, and you pay the model’s price. For a finished ad with a voiceover, the AI ad generator is faster, because it turns one product photo into a UGC ad. If you don’t want to write a prompt, Trends are 7 one-click effects at 50 credits each: upload a photo and get a clip. And cinematic AI video has 33 shot presets (camera moves, looks, moods, scenes), where only Kling 3.0 Pro and Seedance 2.0 work with a starting photo.