Explainer

Can ChatGPT Make a Video? An Honest Answer

No, ChatGPT cannot make a finished, narrated video by itself, and Sora is gone. What it can make (script, storyboard, images, a rough silent clip), what it cannot, and a worked example of a student video where ChatGPT does the half it is good at.
By openCanviz • November 12, 2026

6 min read

No. ChatGPT cannot make a finished video on its own: it will not give you an MP4 with pictures, a voiceover and timing that match. It can write the script, plan the scenes, generate still images, and, through its code tool, sometimes stitch images into a short silent clip. OpenAI's video model, Sora, was shut down in 2026 (the app in April, the API in September) and never arrived inside ChatGPT. Videos that appear "in ChatGPT" today come from connected apps run by other companies, such as HeyGen, Runway or invideo, using your account with them. The practical answer is to let ChatGPT help with the script and use a video tool for the rest.

What ChatGPT can and cannot output

A video has six parts. Here is what ChatGPT does with each one, as of October 2026.

Part of a videoCan ChatGPT make it?The honest detail
ScriptYes, wellFluent and fast. Facts, dates and numbers still need checking.
Storyboard or scene listYesGood at splitting a script into scenes and describing what each shows.
PicturesPartlyIt generates still images one at a time. Keeping the same character or style across ten scenes is hard, and text inside images is often misspelt.
VoiceoverNo, not as a fileVoice mode speaks replies aloud, but it is a conversation, not a narration track you can export and time to scenes.
Animation and timingBarelyThe code tool can make a simple silent clip or GIF from charts or images. No narration sync, no editing.
Finished MP4NoNot without a third-party app connected to the chat.

The row that surprises people is the last one. Many guides from 2025 say "ask ChatGPT to use Sora". That stopped being true when OpenAI discontinued Sora. If you want a full list of the apps that do render video from inside a chat, and what each kind actually produces, see the best ChatGPT video generators and what they actually do.

Why it cannot, briefly

ChatGPT is a language model with some tools attached: an image generator, a voice interface, a code sandbox. Video needs all of those working together on one timeline, where scene four's picture appears exactly when scene four's sentence is spoken. Nothing inside ChatGPT owns that timeline. The code sandbox can technically write frames to a file, which is why you sometimes see a ChatGPT-made GIF of a bouncing ball or a growing bar chart, but it has no voice to sync to and no way to draw a coherent illustrated scene.

The useful split

Since ChatGPT is good at words and poor at the timeline, give it the words.

ChatGPT does: choosing the angle, a first draft of the narration, splitting it into scenes, suggesting what each scene shows, tightening sentences that are too long to say in one breath.

You do: checking every fact, deciding what stays, reading it aloud.

A video tool does: drawing each scene, voicing the narration, timing pictures to words, exporting the file.

The split matters for school work too. If the video is marked, the understanding has to be yours. A script you checked and rewrote is your work in a way a one-click video is not, and how to make a video for a school project covers what teachers tend to look for.

Worked example: a three-minute biology video

Say the task is a three-minute video for class on how vaccines train the immune system. At about 150 spoken words a minute, that is roughly 450 words of narration.

Step 1, the prompt. Something like: "Write narration for a three-minute explainer for 15-year-olds on how a vaccine trains the immune system. About 450 words. Split into seven scenes of two to four sentences. For each scene give NARRATION and one VISUAL line saying what is drawn. No timestamps, no music cues."

Step 2, what comes back (trimmed to the scene list):

  1. A germ the body has never met. The body is slow the first time, which is when people get ill.
  2. What a vaccine contains: a harmless piece of the germ, a weakened or inactivated germ, or instructions (as in mRNA vaccines) for cells to make one harmless piece.
  3. Immune cells notice the unfamiliar piece, the antigen.
  4. B cells make antibodies shaped to fit that antigen. T cells help coordinate and attack infected cells.
  5. Some of those cells become memory cells and stay in the body.
  6. Months later, the real germ arrives. Memory cells recognise it and the response is fast and strong.
  7. Why boosters exist: memory can fade, and some germs change.

Step 3, the check. This is the part ChatGPT cannot do for you. In drafts like this, look for three classic slips: saying a vaccine "contains the virus" without saying it is weakened, inactivated or just a piece; mixing up antibodies (proteins) with the cells that make them; and claiming immunity is lifelong for everything. Fix each against your textbook.

Step 4, the video. Paste the checked narration into a tool that draws scenes from text. Each of the seven scenes becomes one drawing: the germ, the vaccine's contents, the antibody fitting the antigen like a key, the memory cells waiting. Watch it with the sound off; if scene four does not show an antibody attached to an antigen, redraw it.

Make it

  1. 1

    Ask ChatGPT for narration, not a screenplay

    Ask for two to four sentences per scene and one short visual line each. Strip any timestamps, camera directions or music cues it adds anyway.

  2. 2

    Check every claim against a real source

    Textbook, class notes or an official page. Names, numbers and cause and effect are where drafts go wrong.

  3. 3

    Read it aloud and cut

    Any sentence you stumble on, shorten. At 150 words a minute, count the words to know the length before you make anything.

  4. 4

    Paste it into openCanviz

    Choose Keep my wording so the narration is exactly the script you checked, set the length, and pick whiteboard for a process or doodle on paper for something lighter.

  5. 5

    Edit scene by scene

    Fix any label or arrow that is wrong in the editor without regenerating the whole video. Replace the generated voice with your own recording if your teacher asks for it.

If your script is already written, how to turn a ChatGPT script into an animated video has the exact cleanup list and a prompt that produces a paste-ready script.

Common questions

Can ChatGPT make a video from my images? If you ask, its code tool may combine images into a short silent clip or GIF. It will not add narration timed to each image or let you edit the result properly. For a narrated slideshow you need a video tool.

Will Sora come back inside ChatGPT? OpenAI has not announced that. As of October 2026 Sora is discontinued, and any site using the name is not an OpenAI product. Plan around what exists.

Is a video made from a ChatGPT script allowed for school? That depends on your school's AI policy, so read it. Many allow AI help with drafting if you say so and the understanding is clearly yours. Rewriting the script in your own words and checking every fact is the safe version.

Can ChatGPT write a script in Hindi or Spanish for a video? Yes, and it is better to write the script in the language the video will be in than to translate afterwards. Have a fluent speaker read it once, because fluent-sounding mistakes are harder to spot in a second language.

Split the job in two

Use ChatGPT for a 450-word draft, spend twenty minutes checking and rewriting it, then paste it into openCanviz with Keep my wording on. You get a three-minute video where every sentence is one you approved. It is free to start.

Made with openCanviz

Turn any concept into an animated explainer

Type an outline, get a narrated, animated whiteboard video in minutes. No design skills, no timeline scrubbing. Free to start.

Start free
Keep reading
Как превратить конспект в видео для повторения перед ОГЭ и ЕГЭ

Как сделать из своего школьного конспекта короткое видео с озвучкой на русском для подготовки к ОГЭ и ЕГЭ: как разбить его по кодификатору ФИПИ, сколько минут получится, что вырезать, пример по биологии и честный разговор о том, почему одного просмотра мало.

Como transformar seus resumos em vídeo de revisão para o ENEM e os exames nacionais

Um método para transformar o resumo de um conteúdo em vídeo de revisão narrado em português, pensado para a lógica do ENEM e da TRI e para os exames nacionais em Portugal. Com exemplo completo de cenas sobre bioacumulação, um plano de revisão espaçada e os limites honestos do método.

नोटलाई रिभिजन भिडियो कसरी बनाउने: SEE र कक्षा १२ को तयारीका लागि

आफ्नो एउटा पाठको नोटलाई छोटो, अध्याय अध्यायमा बाँडिएको रिभिजन भिडियोमा कसरी बदल्ने, आवाज आफैं किन रेकर्ड गर्ने, र मेन्डेलको वंशाणुको पूरा उदाहरणसहित। साथै भिडियो हेर्नु मात्र पढाइ होइन भन्ने इमानदार कुरा पनि।

오답노트를 복습 영상으로 만드는 법: 내신과 수능 대비

개념 정리 노트와 오답노트를 한국어 내레이션이 들어간 짧은 복습 영상으로 바꾸는 방법. 내신 시험 범위와 수능 영역별로 영상을 어떻게 나누는지, 한국사 예시 장면 목록, 그리고 영상 시청만으로는 공부가 되지 않는다는 솔직한 한계까지 다룹니다. 오답마다 틀린 이유를 장면으로 만들고, 장면 제목을 시험 문제로 바꾸고, 간격을 두고 다시 보는 순서를 단계별로 정리했습니다.


All Rights Reserved.