AI Text to Video Generator

An AI text to video generator makes a video clip from a written prompt alone. You describe who is in the shot, what happens, where it happens and how the camera moves, and the model renders every frame of it. There is no source footage and no starting image: the words are the whole brief.

That makes text to video the fastest way to get from an idea to a moving shot. It is also the hardest kind of prompt to write well, because a still image only needs a subject and a look, while a clip needs a beginning, an action and an end. A good text-to-video prompt reads like a line from a shooting script, not a caption.

Mage puts several text-to-video models behind one prompt bar. Cherry 2 Pro, Mage’s flagship, renders clips of up to 30 seconds at up to 1080p with sound. Lemon is the all-round model at a friendlier price, Kiwi handles everyday clips with nothing to set up, and Melon is built for anime. If you would rather start from a picture, the image to video generator is the sister page to this one.

Start with the action, not the scenery. Models give the most attention to what you describe first, so a prompt that opens with three lines of setting often returns a beautiful, nearly frozen shot. Put the verb up front: a woman sprints across a rooftop, a dragon dives through cloud, a cat knocks a glass off a table.

Then direct the camera. Words such as tracking shot, slow push-in, low-angle, aerial, handheld, whip pan and crash zoom change the clip more than adjectives do. If you want several shots in one clip, write them in order with rough timings, for example a wide shot for the first two seconds, then a close-up on the face, then a reveal of what she is looking at.

Finish with light, lens and sound: golden hour, neon rain, 35mm film, distant thunder, crowd noise. Cherry generates audio with the picture, so a line about sound is part of the prompt, not an afterthought. When you already have a frame you love, the image to video generator keeps it exactly and animates it, which is more predictable than describing it again in words.

• Lead with one clear action and one subject.
• Name the camera move: push-in, tracking, aerial, handheld.
• For multi-shot clips, list the shots in order with timings.
• Describe the sound you want; Cherry renders audio with the clip.
• Draft short and low-resolution, then render the keeper long and sharp.

Length and resolution are the two settings that change the price of a clip the most, so treat them as the last decision, not the first. Find the prompt at 480p and a short duration, then render the final version at 720p or 1080p and the length the scene needs. Cherry 2 Pro takes durations from 4 to 30 seconds; Lemon takes anything from 2 to 30 seconds, one second at a time.

Aspect ratio is about where the clip will live. 16:9 suits trailers, YouTube and anything shown on a screen; 9:16 fills a phone for Reels, Shorts and TikTok; 1:1 sits well in a feed. Cherry renders all three, and Lemon adds 4:3 and 3:4.

A longer story is usually better as several clips than one. Write each shot as its own prompt, keep the same @character in each, and join them in Storyboard, which combines up to six clips into one MP4. For a story that writes itself, see the AI movie generator.