# Seedance 2.5 Text to Video AI > Create a video from text with Seedance 2.5. Describe the subject, what changes during the shot, and how the camera should move. No starting image is required. Web page: https://banana2.pro/seedance2_5_text_to_video_ai ## Current settings and credits - Model: Seedance 2.5 - Price: 480P: 11 credits per second; 720P: 23 credits per second; 1080P: 41 credits per second - Aspect ratios: 16:9, 9:16, 4:3, 3:4, 1:1, 21:9, auto - Duration in seconds: 5, 10, 15, 20, 25, 30 - Text-only input; this entry point does not accept reference images. - Default resolution: 720P Review the total shown in the panel before submitting. The values above come from the current website configuration; examples below explain the workflow. ## From a written scene to a moving shot Seedance 2.5 text to video is useful when you need to visualize an idea before you have footage or reference photography. You can sketch a product reveal, establish a fictional location or test a short visual story. The prompt supplies the scene, while the output settings define its frame and length. Text to video works best when the description separates appearance from action. “A red bicycle outside a bakery” establishes a subject. “A cyclist unlocks it and rides slowly out of frame” adds a sequence that can be seen. Camera instructions then explain whether the viewer stays still, follows the cyclist or moves toward a detail. Treat the first video from text as a test of your brief. If the subject looks right but the action happens too quickly, simplify the sequence or increase its duration. If appearance must match an existing object, switch to an image reference workflow instead of repeatedly trying to describe that exact object in words. ## How to write a Seedance 2.5 video prompt Build a short scene with a clear order of events, then revise one decision at a time. ### 1. Define the scene before adding style Begin a text to video prompt with the main subject, location and visible action. Include time of day when lighting matters. “A baker sets a tray of bread on a wooden counter at dawn” gives the model more usable direction than a list of style adjectives. ### 2. Specify camera and pacing Give Seedance 2.5 one camera instruction, such as a locked wide shot or a slow sideways tracking shot. For longer clips, describe events in order. Avoid asking for a close-up, aerial view and rapid orbit all at once unless the changes are essential to the story. ### 3. Choose a test length and inspect the result Start text to video with five or ten seconds if one action is enough. Pick the aspect ratio for your final placement and review the displayed price. After generation, check whether the opening, action and ending happen in the intended order. ## Text to video prompt examples you can adapt Replace the subject and setting while keeping each example focused on a single visual idea. ### A quiet lifestyle scene Try this Seedance 2.5 prompt: “A person sits at a small kitchen table and opens a notebook. Morning light falls across the page. Locked medium shot, natural movement, faint street ambience.” This video from text brief leaves room for the action without introducing extra characters or cuts. ### A product reveal without a reference For a concept product, use text to video to explore presentation: “A matte blue desk lamp stands in a dark room. It switches on, lighting the wall with a warm circle. Slow push toward the lamp.” Use a real photo later if the final design must match a manufactured item. ### An establishing shot for a story A video prompt can establish place before a character appears: “An empty train platform after rain. A distant train approaches as reflections tremble in shallow puddles. Wide stationary camera.” Keep the weather, lighting and camera consistent if the shot will sit beside other footage. ## Questions about Seedance 2.5 text to video ### How do I turn text into a video with Seedance 2.5? Enter your video prompt, choose the length, resolution and frame shape, then generate. This text to video page does not require an uploaded picture. Describe events that are visible on screen rather than background information the audience cannot see. ### What makes a good text to video prompt? Use subject, setting, action, camera and lighting as a checklist. Give Seedance 2.5 enough detail to choose a composition, but remove instructions that compete with each other. A clear action such as folding a letter is more useful than “make something exciting.” ### How long can a Seedance 2.5 text to video clip be? The available lengths are 5, 10, 15, 20, 25 and 30 seconds. A longer video from text needs a reason to keep moving: an unfolding action, a reveal or a change in the environment. Avoid filling thirty seconds with a one-second event. ### How many credits does text to video use? Seedance 2.5 costs 11, 23 or 41 credits per second at 480P, 720P or 1080P. For ten seconds, that is 110, 230 or 410 credits. The charge shown in the panel uses your chosen length and resolution; paid credits are required. ### Can I request dialogue or background sound? Audio generation is enabled. Include a concise sound direction in the video prompt and check the generated audio before sharing. This text to video interface does not let you upload a recording, select a specific voice or guarantee exact dialogue timing. ### Can text to video keep a character identical across clips? Repeated descriptions can help, but separate generations may change a face, outfit or prop. When identity is important, use the [image to video page](https://banana2.pro/seedance2_5_image_to_video_ai) with consistent reference pictures. Seedance 2.5 prompts alone do not lock an exact character across every result. ### What should I do if a text to video task is delayed? Open your profile to check the existing task. The background recovery process checks delayed generations, so immediately submitting another request can create another paid task. Credits are returned for a confirmed failure, rather than for a temporary delay. ## Write one shot you can picture clearly Use Seedance 2.5 text to video to test a scene before building a longer sequence. Keep the first prompt specific enough to judge what worked. [Open the tool](https://banana2.pro/seedance2_5_text_to_video_ai)