AI Video Generator from Text and Images

An AI video generator creates video from inputs such as a written prompt or a reference image. CiciGPT supports text to video and image to video, so you can start with a scene description or an existing picture.

Text to video starts with a description of the scene you want to create. Image to video starts with an existing visual and uses your instructions to guide motion. Choose image to video when the starting appearance matters, and text to video when you want to explore a new scene.

Upload one JPEG, PNG, or WebP image up to 20MB. Use a photo or illustration you have permission to upload. Choose a clear subject, enough room for the planned movement, and a composition that already resembles your intended shot.

Choose a whole number from 1 to 15 seconds per generation. The default is 5 seconds. A longer clip uses more credits; this workspace does not include a video-extension tool.

CiciGPT offers 480p and 720p video settings. Choose Auto, 16:9, 4:3, 1:1, 9:16, 3:4, 3:2, or 2:3. Select 9:16 for a vertical clip. These are generation settings; exact pixel dimensions depend on the framing. There is no 1080p, 4K, or video-upscaling option in this workspace.

Video generation requires sign-in and paid credits. It has no guest video allowance, and welcome credits cannot be used for video generation. You can watch the public examples and load their inputs before signing in.

Grok Imagine Video 1.5 costs 10 credits per second at either available resolution. A 5-second clip costs 50 credits, a 10-second clip costs 100, and a 15-second clip costs 150. The Generate video button shows the selected cost. Credit-pack prices are shown on the pricing page; the dollar cost of a clip depends on the credit pack purchased.

Credits are deducted after the server creates a job. If the processing provider confirms that the job failed or was cancelled, the backend returns its deducted credits. An initial moderation rejection happens before job creation and deduction. A browser timeout or Stop waiting action does not establish server-side failure, and an unwanted result from a completed job is not a technical failure. If the balance is not restored after a confirmed failure, contact support with the job ID.

Generation time varies with the processing queue and the request. CiciGPT does not promise a fixed completion time. Keep the page open while the job is processing. If you stop waiting, retain the job ID and use Check submitted video again to request its result.

Describe one clear action and the camera behavior you want. Start with a simple shot, review the result, and change one instruction at a time. For image to video, use a clear source image and focus your prompt on motion.

Not always. Generated motion can alter faces, lettering, edges, or the shape of an object. Ask for the details you want to preserve, keep the action focused, and inspect the full clip before using it.

Audio is not a selectable setting in this workspace. Play the generated video and inspect any soundtrack before using it. CiciGPT does not provide a separate voiceover, music, or audio-editing control here.

This workspace has no watermark setting. Inspect the actual generated file before publishing; this page does not promise watermark-free output for every result or plan.

Check the CiciGPT Terms of Service, applicable model-provider conditions, and rights to your source material before commercial use. Permission to upload an image does not establish rights to every person, brand, or other protected element in it. CiciGPT does not guarantee copyright ownership, exclusivity, or acceptance by an advertising platform.

Generating a video sends the image, prompt, and settings to CiciGPT and its processing providers. A local preview is not a generation request. The Privacy Policy explains content handling; a fixed deletion period for every server and processing system is not promised. Avoid confidential uploads and contact support about retention or deletion questions.

No. Stop waiting ends result polling in this browser. A submitted job can continue on the server and may still use credits. Use Check submitted video again to retrieve its result, and keep the job ID if you need support.