Creator Hub
Back

Updated

How to Make AI Videos Directly in ChatGPT (Full Guide)

Higgsfield
14 min
GPT Banner

ChatGPT doesn’t generate video itself, and OpenAI discontinued Sora in 2026. To make an AI video in ChatGPT, connect a video app from the Plugins Directory, such as Higgsfield, Runway, OpenArt, HeyGen or invideo. Sign in to that service if it asks, then describe your shot. The three tables below cover 17 video plugins and which account pays for generation.

Key Takeaways

  • Choose a shot generator such as Higgsfield or Runway for original footage, a presenter app such as HeyGen for an explanation, or a timeline editor such as invideo or Descript for a sequence built from existing media.
  • As of November 2026, this guide covers 17 plugins from the ChatGPT Plugins Directory that generate or edit video, including Higgsfield, Runway, OpenArt, Magnific, HeyGen and invideo.
  • A ChatGPT plan doesn’t pay for video: Higgsfield, Runway, OpenArt, HeyGen and Descript charge plugin generations to your credits with that service.
  • invideo is the exception for editing: edits and exports through ChatGPT are free, while generation, transcripts and paid stock use invideo credits.
  • Installing a plugin and signing in to the video service are two separate setup steps, and a managed workspace may also need administrator approval.
  • A clear prompt names six elements (subject, action, setting, camera, light and mood) and asks for the credit cost before generation.

Can ChatGPT Make Videos Now?

ChatGPT doesn’t generate video on its own: a connected video app makes the clip, and ChatGPT helps write the brief and request changes. OpenAI discontinued the Sora web and app experiences on April 26, 2026, followed by the Sora API on September 24, 2026.

Older instructions that tell you to open Sora or use a ChatGPT subscription to access it describe a discontinued product. Use a current video app from the Plugins Directory instead.

Which Video Apps Can You Connect to ChatGPT?

ChatGPT connects to video apps such as Higgsfield, Runway, OpenArt, HeyGen and invideo through plugins. This guide covers 17 video plugins from the ChatGPT Plugins Directory. They fall into three groups: apps that generate shots from a prompt, apps for presenters, lip sync and dubbing, and apps that assemble or edit a sequence. Account requirements and billing vary by app: most plugins need a connected account with the service, while Explain Video Generator lists no login. In the tables, MCP means the connection a plugin uses to reach the service. The tables show those rules; they don’t rank video quality.

Apps That Generate Shots From a Prompt

The nine plugins in this table create new footage from text or reference images.

Apps That Generate Shots From a Prompt

App

What You Can Make

Account Required

How You Pay

Higgsfield AI

Generated scenes from text or reference images, product clips, video ads

Higgsfield account with an active subscription

Higgsfield credits at standard rates; Unlimited and free generations don’t apply in ChatGPT

Runway

Generated shots, image animation, edits to existing footage, multi-shot story videos

Runway account

Runway credits, the same as in the Runway app; Unlimited Mode and Explore Mode aren’t supported; Runway MCP is listed on the Pro and Max plans

OpenArt

Text-to-video shots and animated images with models such as Seedance 2.5 and Kling 4.0

OpenArt account

OpenArt credits; MCP access is listed from the Starter plan

Magnific

Images, video and audio; ads, UGC videos, faceless videos

Magnific account on a paid plan

Every generation uses credits, including models that are unlimited on the web plan

Krea

Images, video and 3D assets

Krea account

Krea plans come with compute units; the listing doesn’t say how plugin generations are billed

fal

Images, video, audio and 3D assets

fal account

Pay per model run at API prices; the MCP server itself is free

Picsart

Images, video and audio; short videos from recordings and designs

Picsart account

Picsart credits, separate from your ChatGPT plan

Creative Claw

Videos, product ads, images and voiceovers

Creative Claw account

Pay-as-you-go credits, $10 for 1,000; 100 free credits at sign-up

AI Video Maker (InfoseekAI)

Short videos with Seedance 2 and 2.5, video edits and extensions

Infoseek account

Usage-based, from a prepaid balance; no subscription

Apps for Presenters, Lip Sync and Dubbing

The three plugins in this table make a person or character speak, or move speech into another language.

Apps for Presenters, Lip Sync and Dubbing

App

What You Can Make

Account Required

How You Pay

HeyGen

Avatar explainers, narrated product demos, HeyGen Video 1 clips, dubbing into 175+ languages

HeyGen account (a free account works to start)

HeyGen account credits; new accounts get free credits to start

VEED Video Generator

Videos with a chosen character and voice

Not stated in the listing

Not stated in the listing; see VEED plans

sync. labs

Lip sync, talking photos, dubbing, word fixes in recorded speech

sync. labs account

Monthly fee plus per-second usage

Apps That Assemble and Edit Sequences

The five plugins in this table build a video from a script or your own footage and let you edit the cut.

Apps That Assemble and Edit Sequences

App

What You Can Make

Account Required

How You Pay

invideo

Videos assembled from scripts or footage, with narration, music, captions and generated B-roll

invideo account (you can create one during sign-in)

Edits and exports are free; generation, transcripts and paid stock use invideo credits

Descript

Text-based editing of video and audio, captions, exports to Premiere Pro or DaVinci Resolve

Descript account

AI credits from your Descript plan; no separate charge for MCP

Visla Video Maker

Narrated videos with stock footage and music from a ChatGPT script

Visla account

Not stated in the listing; see Visla plans

Explain Video Generator (Scrimba)

Narrated explainer videos with diagrams, animations and code walkthroughs

No login required

Free to start, per the listing

Manus

An agent that can produce a video along with slides, websites and research

Not stated in the listing

Not stated in the listing; see Manus plans

Directory listings and billing rules checked on November 13, 2026. Availability and plan details may change. We did not run a same-prompt quality test.

The directory also lists Kling AI for Terminal, but it runs through Codex, not a regular ChatGPT chat, so it isn’t in the table.

For the tennis shot later in this guide, choose an app that generates footage, such as Higgsfield. Higgsfield is a generative platform for creative production with 35+ frontier models for generating images and turning them into video scenes. For an explanation delivered by a presenter, start with HeyGen. For a sequence built from your recordings, consider invideo: its ChatGPT workflow opens your project on a timeline you can review and adjust in the browser.

What Do You Need Before You Start?

You need a ChatGPT account that can use your chosen app, a connected account with that video service if the app requires one, and access to the feature you plan to use. The first decision is the output: one shot from Higgsfield, Runway or OpenArt, a presenter explanation from HeyGen, or an edited sequence from invideo or Descript. Each needs a different brief.

  • Choose a destination format, such as vertical 9:16 for a TikTok or Instagram Reels clip or horizontal 16:9 for a YouTube video or a presentation.
  • Prepare any reference photo or footage you want the app to use. Say whether a photo supplies the subject’s appearance or the opening frame. Use only photos, footage and voices you have the right to use, and get a person’s permission before you put their face or voice in a video.
  • Check the service’s credit balance before asking for several outputs. Installing a plugin doesn’t give you access to every model or feature: Runway, for example, lists Runway MCP on its Pro and Max plans.

Installing the plugin in ChatGPT and signing in to the video service are separate steps. In a ChatGPT Business, Enterprise or Edu workspace, an administrator may also need to approve the plugin.

How Do You Connect an App and Make Your First Video?

Connect a video app through the Plugins Directory, sign in to its service if asked, and start a new chat. Then name the app and describe one output. The six steps below go from finding the app to reviewing a first clip, with a cost check before any paid generation.

Step 1: Find Your Video App

Select the Plugins icon in the ChatGPT sidebar, type the service’s name in the Search plugins box and open its page. Read the listed capabilities to confirm that the app supports the output you want.

Step 2: Install the Plugin

Select the plus button next to the plugin in the list, or select Install plugin on the plugin’s page.

Step 3: Authorize the Service Account

If the app asks you to connect, follow the prompt and sign in to the video service. Authentication may appear during installation or when you first use the app. Sign in to the account whose credits you want to use. The detailed Higgsfield setup is covered in ChatGPT with Higgsfield.

Step 4: Describe One Video

Start a new chat, type @, and select the installed app. Give the subject, action, setting, camera, light, mood, length and format. If you have a reference image, attach it and explain its role. The tennis prompt below is ready to adapt.

Example in ChatGPT with the Higgsfield plugin. Unlike the text prompt below, the prompt in the screenshot also uses a character reference and a location reference.

Step 5: Review the Proposed Generation

Ask ChatGPT to state the video model, the settings and the credit cost before it starts. Approve one version when the proposed output matches your brief. If a setting isn’t available, pick one of the options ChatGPT offers before it generates. Approve one clip at a time until the first one looks right.

Step 6: Review the Video

Play the result and compare it with your brief. Check the action and camera movement, then inspect faces, hands, text and product details. Download the result or open the service’s project link, depending on the app. Give the next instruction in the same chat so ChatGPT knows which clip you mean.

The finished 10-second clip from the Higgsfield plugin, returned in the same chat.

How Does Payment Work for Video Apps in ChatGPT?

Your ChatGPT plan pays for ChatGPT, and the connected video service bills generation separately. Higgsfield, Runway, OpenArt, HeyGen and Descript charge plugin generations to the credits of your account with that service, while invideo charges for generation, transcripts and paid stock but not for edits. Check the payment rule for your app in the tables, and ask for the cost before you approve a generation.

A credit is a service’s billing unit. The same number of credits on two platforms doesn’t mean the same price or the same amount of video. Costs vary by model, duration, resolution and other settings, so ask for the total charge of the complete request, including any images or audio created along the way.

For more detail on how generation settings affect spend, see AI Video Credits Explained.

How to Write a Video Prompt That Gets the Shot You Want

A good video prompt names six elements: subject, action, setting, camera, light and mood. Add the intended length and orientation, then request the credit cost before the app begins making the clip.

How to Write a Video Prompt That Gets the Shot You Want

Element

What to Specify

Example

Subject

Who or what is on screen

A tennis player in athletic gear

Action

What happens during the clip

He hits a powerful serve

Setting

Where and when

A dramatic canyon tennis court at golden hour

Camera

Shot size and movement

Starts on a close-up of the player’s racket and face, then slowly pulls back

Light

Direction, color and quality

Dramatic golden hour light from the side

Mood

The feeling the scene should carry

Intense and epic

A Copyable Video Prompt Template

“@[app name] Generate a [length]-second [vertical or horizontal] video of [subject] [action] in [setting]. Camera: [shot size and movement]. Light: [direction and quality]. Mood: [feeling]. Before generating, show me the credit cost and wait for my confirmation.”

From vague to specific. A vague request leaves every choice to the model:

“Make a video of a tennis match.”

A more specific prompt gives the model clearer direction.

The Tennis Example

"@Higgsfield Generate a 10-second vertical video of a tennis player hitting a powerful serve in an epic canyon court at golden hour. The camera starts on a close-up of the player's face and racket, then slowly pulls back to reveal the vast canyon landscape. Dramatic golden light from the side, intense mood. Before generating, show me the credit cost and wait for my confirmation.”

To pick the model yourself, add its name to the request, for example “Use Seedance 2.5.” Confirm that the chosen app offers it. For a photo as input, see How to Turn a Photo Into an AI Video.

How to Refine Your Video After the First Generation

Refine the first video by naming the change and what should stay the same. Compare the result with the brief in three areas: motion, composition and pace. A follow-up may need another generation or an edit, if the app supports one, so ask whether the app will generate a new clip or edit the current one, and what it will cost.

  • Motion: Does the action read correctly, and does it run at the right speed?
  • Composition: Are the framing, the subject’s position and the background what you described?
  • Pace: How fast does the camera move, and how does the clip build across its length?

Also look at the details that tend to slip, such as hands, text and logos.

A Follow-Up for the Tennis Shot

Say the first generation of the tennis prompt came back with the camera pulling back too fast and the player's expression unclear. The next message can name both fixes and what to keep:

"Keep the same scene and light. Slow the pull-back so it takes the full ten seconds. Make the player's face clearer and more expressive. Keep the canyon landscape visible.”

A Follow-Up for the Tennis Shot

What to Change

Example Follow-Up

Pacing

“Slow the camera pull-back to the full ten seconds”

Composition

“Frame the tennis player on the left third with the canyon behind”

Motion

“Have the player swing slowly with perfect form”

Color or light

“Make the light more dramatic and golden”

One object or the background

“Replace the canyon with a desert court at sunset”

Download a version you like before you ask for another. For a longer sequence, develop the shots first with How to Make an AI Video From a Script.

What Mistakes Should You Avoid?

Avoid requests that leave the app, the output or the charge unclear. For a first video, keep the task to one shot or one short edit, then review it before you ask for more.

  • Name the app with an @ mention, for example @Higgsfield. Without it, ChatGPT decides by itself whether an installed plugin fits the request.
  • Decide whether you need one generated shot or a finished video with several scenes, narration and captions.
  • Ask what the request will cost. In invideo, “edit this” can mean a free timeline change or a paid generation.
  • Request one specific revision at a time, and state which visual details to keep.

Which ChatGPT Video App Should You Choose?

Choose a ChatGPT video app by the result you need: ChatGPT has no built-in video model, so the plugin decides the models, the account you need and how you pay.

  • Generated shots and ads from a prompt: Higgsfield, Runway, OpenArt or Magnific.
  • A presenter, lip sync or dubbing: HeyGen or sync. labs.
  • A sequence built from your own footage or a narrated explainer: invideo, Descript, Visla or Explain Video Generator.

Whichever app you pick, connect its account if it needs one, write one clear brief, check the charge before generation and use specific follow-ups for the next version.

How to Make AI Videos Directly in ChatGPT (Full Guide)

Connect Higgsfield to ChatGPT

Got any questions left?

by Higgsfield