> For the complete documentation index, see [llms.txt](https://docs.arcade.software/kb/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.arcade.software/kb/ai-video/ai-video.md).

# AI Video

Quickly create on-brand teaser videos, sizzle reels, and product overviews. Easily tweak headlines and voiceovers to stay on message.

## Overview

AI Video turns your Arcade demos, brand kit, and a short description into a polished, on-brand video — automatically.

Video generation is conversational. Instead of filling out a form, you describe the video you want to Arcade's AI assistant, and it builds a complete first draft (typically in under a minute). You then keep refining the video by chatting with the assistant or by editing directly on the timeline. Big changes — shortening the intro, swapping a visual, reordering scenes — can be made either way.

Your video is built from scenes on a timeline — see [Scenes](/kb/ai-video/ai-video/scenes.md) for what they are and how to add one. For what you can change by chatting versus editing directly, see [Editing](/kb/ai-video/ai-video/editing.md).

Arcade also remembers preferences you set or demonstrate, so you don't re-explain your style on every video. See [Memory & Preferences](/kb/ai-video/ai-video/memory-and-preferences.md).

***

## Creating a video

### How do I create a video?

You can start in three ways:

* From the Create button on your home screen (team or personal workspace)
* From an existing Arcade — click Share, then Generate with AI → Video
* From any webpage, using [Producer](/kb/ai-video/ai-video/producer.md) in the AI Video Chrome Extension — point it at a page and it drafts a brief for you

You can also start from a [Template](/kb/ai-video/ai-video/templates.md) instead of a blank prompt.

### How does creation work?

You create a video by describing it in a chat with Arcade's AI assistant. In the composer you can:

* Describe the video — your goal, audience, key message, or a full script
* Pick a brand kit — so the video matches your colors, fonts, logo, and voice from the start
* Pick a quality tier — `Light` (Gemini 3.8, fastest), `Medium` (Sonnet 5.5, balances speed and quality), or `High` (Opus 5.5, maximum detail and polish)
* Set a length — `15s`, `30s`, `45s`, `1m`, `1m30`, or `2m` from the length dropdown; see [Plans, Credits & Pricing](/kb/ai-video/plans-credits-and-pricing.md) for how length and model affect credit cost
* Set an aspect ratio — `16:9`, `1:1`, or `9:16`
* Attach source material — other Arcades, URLs, and files. Supported file types include images (.jpg, .jpeg, .png, .svg), PDFs, Office documents (.docx, .pptx, .xlsx), text and code files (.txt, .md, .json, .html, .csv, and more), and videos (.mp4, .mov, .webm)
* Type @ to pull in an attachment or a video template directly — anything you attach this way shows up as an @ reference in your prompt

If you're not sure where to start, the composer offers a few starter prompts you can click and edit.

### What happens after I generate?

Arcade analyzes your input, applies your brand kit, and produces a full first draft — usually in under a minute. While it works, the chat shows you what's happening — the current phase, the narration it's writing, and other activity — instead of a blank wait. When it's ready you'll see:

* A preview player (play / pause, mute)
* A timeline of all the scenes
* A way to rate the result (thumbs up / down)

Arcade will also flag anything worth double-checking directly in the chat — for example, if source images may carry licensing restrictions, or if it hasn't verified audio levels yet. Treat these as a checklist before you publish, not just a summary.

### Why is the Generate button disabled?

Usually one of two reasons:

* You haven't described the video yet, or
* You're out of AI credits

Generating a video uses AI credits, and so does asking the assistant to edit it afterward through chat — editing directly on the timeline or in the Design panel yourself stays free. Check your balance in the top left or in Settings.

***

## Scenes

Every video is a sequence of scenes on the timeline. Add one from your media library, an upload, a prompt to AI, or Arcade's curated scene library of pre-built text animations, UI-style moments, and a logo reveal.

See [Scenes](/kb/ai-video/ai-video/scenes.md) for the full breakdown, including every scene in the curated library and how brand styling is applied automatically.

***

## Editing your video

After your first draft is ready, you can keep refining it by chatting with the assistant or by editing directly on the timeline and in the Design panel — the chat stays open the whole time you're working, so you can mix both as you go.

Manual edits on the timeline and in the Design panel are always free. Asking the assistant to change something through chat costs credits, at the same per-model rate as generating the video — see [Plans, Credits & Pricing](/kb/ai-video/plans-credits-and-pricing.md). A few chat actions — regenerating a voiceover, fitting music to a new cut, captions, and voice suggestions — don't cost credits either way.

See [Editing](/kb/ai-video/ai-video/editing.md) for the full breakdown of what's available each way, including voiceover, clip controls, background music, captions, and style references.

***

## How your brand kit is used

Your brand kit is the foundation of every generated video. When you pick a brand kit at creation, Arcade pulls from it automatically:

* Colors — applied throughout the video and offered as defaults in the editor's color pickers
* Fonts — your brand typography is embedded in the video
* Logos and assets — used as visual material where they fit
* Voice and tone — shapes the AI-written script and the voiceover's style

Changing a brand kit affects future generations, not videos you've already made. See [Brand Kit](/kb/ai-video/brand-kit.md) for the full breakdown — what's inside a Brand Kit, how to create, edit, and update one, and how it powers AI Video specifically.

Correcting a color, font, or logo while chatting about a video writes that fix back to the Brand Kit itself, so later videos get it right too.

***

## Memory & your preferences

Arcade remembers preferences you set or demonstrate — your preferred voice, music, pacing, length, motion, sound effects, and how your videos open and close — and applies them automatically to future videos. See [Memory & Preferences](/kb/ai-video/ai-video/memory-and-preferences.md) to learn what's remembered, how to review it, and how to turn it off for a single video.

***

## Sharing & exporting

Publish your video to get a shareable link and an exported MP4, manage your generated-video history, and understand watermark behavior. See Sharing & Exporting for the full breakdown.

***

## Other ways to create a video

Beyond the composer on your home screen, you can start or create an AI Video from:

* **MCP** — ask Claude, ChatGPT, or another connected AI agent to create a video for you. See [MCP](/kb/ai-video/mcp.md) for supported clients and setup.
* **Claude Connector** / **ChatGPT Plugin** — install from Resources → Apps to create videos directly from those chat interfaces.
* **AI Video Chrome Extension** — see [Overview](/kb/ai-video/ai-video/overview.md) and [Producer](/kb/ai-video/ai-video/producer.md).

***

## A note on older videos

Videos created before the chat-first editor may open in an earlier editor. The newest editor now supports manual timeline editing too — reorder, trim, add, duplicate, copy/paste, and delete scenes — alongside chat. The older editor additionally offers pan & zoom, per-section background, transition, and animation settings, and GIF and multiple-resolution export, which aren't in the new editor yet. Those extra controls apply only to those older videos.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.arcade.software/kb/ai-video/ai-video.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
