Skip to main content

Video MCP Server

Ask Claude or ChatGPT for a video.
Get an MP4 back.

Add iArt to Claude or ChatGPT once and sign in, with nothing to install and no code. Describe the video in the chat; iArt makes the motion graphics video and sends back an MP4 and a share link.

Beta · Pro, Ultra and Max plans

Made by Claude Code through the iArt MCP server from one request in a repo; the full session is below.

What is a video MCP server?

Video, as a tool
your agent can call.

MCP (Model Context Protocol) is how AI assistants use outside tools. A video MCP server gives the assistant a tool that makes videos. iArt's is a remote MCP server: add one URL to Claude, ChatGPT or another AI tool, sign in, and it can start a motion graphics video, check on it, revise it and fetch the result, in the same chat where you are already working.

The assistant does the part it is good at: it reads what you've shared in the chat, such as a doc, your notes, a web page or your logo, and writes a detailed brief with the real names and numbers. iArt does the rest: it writes the animation, adds voice-over if you want it, and renders it. Back come an MP4, a share link anyone can watch, and a link that opens the project in the iArt editor.

It makes motion graphics: kinetic typography, explainers, logo reveals, data animations, product and launch videos. You can attach your logo, screenshots or a PDF to the chat and ask for them in the video.

For developers: setup for Claude Code, Codex, Cursor and VS Code is in the MCP docs; endpoints, fields and errors are in the API reference.

A real session

One request, from a repo to a video.

We opened Claude Code in a folder holding our open-source motion-skills README and our logo, connected the iArt MCP server, and typed one request. Below is what happened, step by step, and the video it made.

  1. You

    Make a 15-second 16:9 launch video with iArt announcing what this repo is: the open-source motion skills, with the real numbers from the README. Use our logo.svg. Wait until it's done and give me the links.

  2. Tool call

    Bash ls -la && cat README.md && head -c 3000 logo.svg

  3. Tool call

    upload_file name: "logo.svg" → upload link (sent with curl)

  4. Tool call

    create_video aspect_ratio: "16:9" files: [logo.svg] prompt: "A 15-second 16:9 launch video announcing "Motion Graphics Skills for AI Coding Agents" … Scenes (exact on-screen text): 1. (0–2.5s) logo.svg reveals and settles; text: "Now open source". 2. (2.5–6s) Headline … 3. (6–10s) Big counting numbers: "56 skills" · "19 packs" · "40+ agents" … 5. (12.5–15s) End card: logo.svg, "npx skills add iart-ai/…" … Voice-over: "Introducing Motion Graphics Skills for AI coding agents …""

  5. Claude Code

    The video is being generated. I'll keep checking its status until it's done.

  6. Tool call

    get_video → status: "writing"

  7. You

    Is it done?

  8. Claude Code

    Yes, it's done: - Share page: https://app.iart.ai/share/bba8bfbb0e20448b18ee8 - MP4 download: (a link that stops working after 24 hours, so download it soon) - Edit in iArt: a link that opens the project in the editor It cost $0.60. I haven't watched it myself. If you want anything changed, tell me what and I'll send a change request to iArt.

The video from this session: Claude Code wrote the brief from the README, uploaded the logo, and iArt made and rendered it, voice-over included.

How to connect

Three steps, no code.

  1. 01

    Add iArt

    In Claude, add a custom connector; in ChatGPT, add a custom MCP server. Paste the URL below.

  2. 02

    Sign in

    Sign in to your iArt account and click Allow. There is no API key to copy.

  3. 03

    Ask for a video

    Describe the video in the chat and mention iArt. A few minutes later you get an MP4, a share link and a link to edit it in iArt.

iArt MCP server URL

https://app.iart.ai/api/mcp

Step-by-step guides with screenshots: set up iArt in Claude and set up iArt in ChatGPT.

Using Claude Code, Codex, Cursor or VS Code? Add the same URL and sign in, or use an API key for scripts. See the developer setup.

What people ask for

  • A video from a doc or a page

    Paste a doc, a PDF or a link into the chat; the assistant picks what matters and briefs an explainer with voice-over.

  • Your logo, your words

    Attach your logo and screenshots; they go into the video along with your product names and numbers.

  • Turn the chat into a video

    You've already worked out a launch, a lesson or a pitch with the assistant. Ask it to make the video from that conversation.

  • Change it in words

    Not quite right? Say what to change; the assistant revises the same video and brings back the new one.

Plans, pricing and limits

The MCP server and the REST API are part of the Pro, Ultra and Max plans. Videos use your account's credits at the same prices as videos made in the app; what one costs depends on the model, the length and how much the agent writes.

PlanVideos at onceHighest resolution
Pro31080p
Ultra64K
Max154K

Good to know

  • –It is in beta.
  • –A video takes a few minutes; your agent checks on it until it is done.
  • –The MP4 link works for 24 hours, so download it. The share link does not expire.
  • –A video is refused up front if your credits can't cover it.
  • –Each key allows 60 new videos or revisions and 300 status checks a minute.

Want something free? The open-source motion graphics skills teach your agent to write and render animations itself.

FAQ

Video MCP Server FAQ

Can't find what you're looking for? Reach out to our team.

Is there an MCP server that makes videos from a prompt?

Yes. iArt runs a remote MCP server at https://app.iart.ai/api/mcp. Connect it to Claude, ChatGPT, Claude Code, Codex, Cursor or VS Code, ask for a video, and the agent gets back an MP4, a share link and an editor link a few minutes later. It makes motion graphics videos: kinetic typography, explainers, logo reveals, data animations and launch videos.

Can Claude or ChatGPT make videos with it?

Yes. In Claude (claude.ai or the desktop app) add iArt as a custom connector; in ChatGPT add it as a custom MCP server. Both sign you in to iArt, so no API key is needed. Then ask for a video in the chat. The iArt help center has a step-by-step guide for each.

Do I need an API key?

No. In Claude and ChatGPT you sign in to iArt when you add it, and Claude Code and Codex can sign in the same way. An API key, created in the iArt app under API, is for scripts, automations and tools that can't sign in.

How much does it cost?

It is part of the Pro, Ultra and Max plans. Videos use your credits at the same prices as in the app: the model's tokens plus $0.09 per rendered scene and $0.045 to combine them. The cost depends on the video's length and the model; the API reference lists the rates.

What does the agent get back?

An MP4 (720p by default, 1080p on Pro, up to 4K on Ultra and Max), an unlisted share link anyone with the link can watch, and a link that opens the project in the iArt editor. The MP4 link works for 24 hours.

Is there a REST API too?

Yes. The same tools are available as a REST API at https://app.iart.ai/api/v1, for scripts and automations. See the API reference.

Make your next video
in the chat.

Add iArt to Claude or ChatGPT, sign in, and ask for your first video.