Describe a video. Prism plans every shot and makes it.
Prism is an AI agent that works like ChatGPT, Claude or Gemini. Ask it for a video and it plans the characters, the scenes and the sets, writes a prompt for each shot and makes them with Seedance 2.5, Veo 3.1 Fast, Gemini Omni 1.1 and H3 Max Turbo.
How the AI video generator works
01Describe the video
Write what happens, or paste a script. You can attach photos, clips and audio as references, or a link to your product page.
02Read the plan
Prism lists every shot with its model and the prompt it will send. Plan mode waits for your approval before it spends credits.
03Prism makes the shots
Prism runs each shot and looks at each result.
04Change one shot
Ask for a change to any shot. Prism remakes that shot and keeps the rest of the video.
Video models in Prism
Seedance 2.5
ByteDance’s video model. It makes shots of 4 to 30 seconds at 480p or 720p with dialogue, sound effects and music. It takes up to 30 reference images, plus reference clips and audio.
Veo 3.1 Fast
Google’s video model. It makes 4, 6 or 8-second shots at 720p or 1080p with audio, from a prompt or from first and last frames.
Gemini Omni 1.1
Google’s video model with native audio, at up to 4K. It also edits a video you supply.
H3 Max Turbo
MiniMax’s fast video model, for 5 to 15-second shots from first and last frames. Prism’s talking heads run on it.
Motion design
Animated graphics written as code: kinetic type, logos, product cards, app walkthroughs and charts, from 3 to 60 seconds long.
What people make with it
Product videos
Attach a photo of your product, or give Prism your product page and it fetches the photos. AI avatars can’t use a product convincingly, so Prism has the creator hold your product, show it up close and describe it. This style is built to follow the ad guidelines of Meta, Google and most other major platforms. In most generations, products with detailed text, such as book covers and skincare labels, render correctly even as the creator turns the product or moves it in and out of frame. We tested our prompts across many generations so this holds up. Veo 3.1 Fast and Gemini Omni 1.1 make 1080p video, which keeps the small text and detail on books and skincare sharp.
Short dramas
Prism’s scriptwriting skill plans the story for social feeds: a hook in the opening shot, tension between the characters and a cliffhanger at the end. Each line is written to keep viewers watching to the last shot. We tested the skill on short-form videos that went viral on TikTok, and it matches the pacing and style of series on ReelShort.
UGC videos
Prism casts from a library of realistic UGC avatars, men and women from their early 20s to their mid-40s. Each one sits in a real-looking room with wall hangings, food on the counter or a dorm-room desk, so the video reads as filmed by a real person.
Music videos
Prism combines avatars and filmed scenes with motion graphics, such as a dancer on a rooftop cut with animated lyrics. Attach your track as a reference.
Explainers
Prism’s motion design skill writes animations in JavaScript: titles, diagrams and charts. Prism cuts them between talking-head shots, and each avatar speaks in its own ElevenLabs voice.
Animation
Prism’s animation skills hold one style across every shot: 3D animation, a hand-drawn 2D feature film, or felt-and-clay stop-motion figures.
Describe the first shot
Frequently asked questions
What is the AI video generator?+
Planegraph’s AI video generator is Prism, an AI agent for images and video that works like ChatGPT, Claude or Gemini. You describe a video, and Prism plans the characters, scenes and sets, writes a prompt for each shot and makes them with models such as Seedance 2.5, Veo 3.1 Fast, Gemini Omni 1.1 and H3 Max Turbo.
Is the AI video generator free?+
Yes, you can start free. Every new account gets 10 free credits and 10 Prism messages a month. Paid plans start at $19 a month and include every feature.
How much does Planegraph cost?+
You can start free: new accounts get 10 credits and 10 Prism messages. Paid plans start at $19 a month with Starter, which gives 1,000 credits a month, enough for about 9 Seedance 2.5 videos or 142 Nano Banana 2 images, and 100 messages. Creator is $29, Pro is $49 and Max is $99 a month, each with more credits and messages.
Which features come with the $19 plan?+
Every feature and every model. Starter at $19 a month includes every model, every advanced feature, the ad library of thousands of real ads, ad templates and the library of realistic AI creators. Higher plans add credits, messages and storage.
Can I use the videos commercially?+
Yes. You own the videos you make with Planegraph, and you can use them in ads, on your channels and in client work.
Does it add captions?+
Yes. Open the video in Planegraph’s video editor and select Add captions. The editor writes captions from the speech in the video, and you can restyle them before you export.
Which formats and sizes can I make?+
Prism makes video at 9:16, 16:9, 1:1, 4:3, 3:4 and 21:9 with Seedance 2.5, and at 9:16 and 16:9 with Veo 3.1 Fast and Gemini Omni 1.1. Motion graphics come in 9:16, 1:1, 4:5 and 16:9. Resolution runs from 480p to 1080p, and up to 4K with Gemini Omni 1.1.
How long can a video be?+
Seedance 2.5 makes shots up to 30 seconds long, and motion graphics run up to 60 seconds. Prism can plan a longer video as several shots and run them in one plan.
Will the text on my product render correctly?+
In most cases, yes. Products with detailed text, such as book covers and skincare labels, render correctly even as the creator turns the product or moves it in and out of frame. We tested our prompts across a wide range of generations so this holds up. If one comes out wrong, ask for a change and Prism remakes that shot.
Do the videos have sound?+
Yes. Seedance 2.5, Veo 3.1 Fast and Gemini Omni 1.1 generate audio with the video, including dialogue, sound effects and music. Talking heads speak in the avatar’s ElevenLabs voice.
How do I write a prompt for an AI video?+
You describe the video in plain words. Prism has a prompt-writing skill for each model built in, so you skip the step of asking ChatGPT or Claude to write a prompt first.
How does Prism write its prompts?+
Prism’s prompts are tested over 100 times to cut artifacts in the video. They give visual direction in descriptive language: where the camera is, what fills the frame and how each person moves. They leave out metaphors and general words, which a video model can read the wrong way.
Is it for beginners or experts?+
Both. Prism can start from a simple prompt and any images or website links you attach. Experts can revise each shot’s prompt in the plan before it runs, or open the finished video in the timeline editor.
Does it have AI actors?+
Yes. Planegraph has a library of realistic AI actors: men and women from their early 20s to their mid-40s, each in a real-looking room such as a kitchen, a bedroom or a dorm. They speak your script in their own voice.
Can I start from my own photos or clips?+
Yes. You can attach photos, videos and audio to your message. Prism uses them as references, such as a product photo for every shot or a clip for the camera move.