AI & MCP · beta

Tell your AI what to make. Doodaro draws it.

Ask your AI agent for a video in plain words. It writes the scenes, picks the style and narrator, and uses Doodaro to draw, narrate and render it. You get an MP4 link, and the project waits in your dashboard if you want to fix a scene.

Connect your agent Automate daily videos

Works today with Claude Code, Codex and Cursor. ChatGPT and Claude chat are coming soon.

Claude Code · doodaro connected

Make a 60-second vertical video about why the sky is blue. Color Chalkboard style, a deep narrator. Show me the plan before you render.

Plan: 6 scenes, about 60 seconds, portrait 9:16, Color Chalkboard, a deep English narrator from the voice library. Generate it?

Yes, go.

list_stylesColor Chalkboard
list_voicesdeep English narrator
create_project6 scenes
generate_assets6 pictures, 6 narrations
render_videocomplete

Done. media.doodaro.com/…/why-the-sky-is-blue.mp4

Open it in your dashboard to redraw or revoice any scene.

Example conversation. Wording and steps vary by agent.

Works with

Use the agent you already have.

Available now

Claude CodeMCP with a key header, or Copy for AI
CodexMCP with a key header, or Copy for AI
CursorMCP with a key header, or Copy for AI
Any MCP clientIf it can send an API key header
Your code and automationsPlain HTTPS: cron, GitHub Actions, n8n, Make, Zapier

Coming soon

ChatGPTNeeds a sign-in connector. Coming soon.
Claude chat appNeeds a sign-in connector. Coming soon.

Two ways to connect

Paste a text, or add a URL.

Both use the same Doodaro API and the same keys. Create a key in Developer or click Copy for AI in AI Connect, both in your dashboard.

One session

Copy for AI

Click Copy for AI. The dashboard makes a new key and copies it with the full agent guide. Paste it into your coding agent, then ask for a video.

paste into your agent

# Paste into Claude Code, Codex or Cursor
# (the text you copied from AI Connect, then your request)

# Doodaro API access
Base URL: https://app.doodaro.com/api/v1
API key: ddr_live_...   (secret, revoke it in Developer)
...agent guide...

Make a 60-second vertical video about why the sky is blue.
Color Chalkboard style, a deep narrator. Show me the plan first.

How Copy for AI works →

Every session

Doodaro MCP

Add https://app.doodaro.com/api/mcp with your key header. Every API action becomes a tool your agent can call. Codex and Cursor configs are in the setup guide.

Claude Code

claude mcp add --transport http doodaro https://app.doodaro.com/api/mcp \
  --header "Authorization: Bearer $DOODARO_API_KEY"

Setup for Claude Code, Codex and Cursor →

What your agent can do

The same studio, as tools.

Each card maps to real Doodaro API operations. Anything the agent makes opens in your dashboard.

Browse styles and voices

Find the 37 styles and your own templates, and search the voice library by language or description.

list_styleslist_voices

Create a project from scenes

Write the title, narration and picture description for every scene, then create the project in one call.

create_project

Edit one scene

Change the narration, the picture description or the voice of a single scene. Add or delete scenes.

update_sceneadd_scenedelete_scene

Redraw or revoice one scene

Generate pictures and narration for the whole video, or only for the scenes you name.

generate_assets

Render the MP4

Queue the full video, or a preview of one scene, on the project's pinned engine.

render_videolist_renders

Check progress

Read any job until it finishes, then return the video link and the dashboard link.

get_jobget_project

Faceless YouTube

Run a faceless YouTube channel with your AI.

A realistic week. The agent does the repetitive production work. You make the choices and check the output.

1
Once
Set the channel

Pick a niche, one drawing style and one narrator. Write a short channel brief your agent reads every time.
You
2
Weekly
Fill the calendar

List the next 7 to 14 topics in a sheet or CSV, with the format and platform for each.
You, with help from your agent
3
Weekly
Draft the scripts

The agent writes each script and the scene-by-scene picture descriptions.
Your agent
4
Weekly
Review

Read every script. Fix facts, cut filler, mark each topic approved.
You
5
Daily or in one batch
Build and render

The agent creates the project, generates pictures and narration, renders, and returns the MP4 link.
Your agent + Doodaro
6
Daily
Post

Upload in YouTube Studio, or let your own automation hand the MP4 to the YouTube Data API or a scheduling tool.
You or your third-party tools

Doodaro does

Draws every scene in the style you picked, stroke by stroke
Records narration with the narrator you picked
Renders the MP4 in 16:9, 9:16 or 1:1
Keeps every project in your dashboard so you can fix one scene by hand

You do, or your own tools do

Choose the niche, the topics and the channel voice
Review every script before it becomes a video
Upload or schedule posts in YouTube Studio or a third-party tool
Turn on platform auto-captions and the AI-content label when asked

YouTube asks creators to avoid repetitive, mass-produced content; review every script and vary your videos. Turn on the platform's AI-content label when it asks. Doodaro does not post or schedule for you yet; Doodaro Publish is coming soon.

Automation

Want a new video every morning?

Set it up yourself with a scheduler you already use. A cron job, a GitHub Actions schedule, n8n, Make or Zapier wakes up, your agent or script builds today's video through the Doodaro API, and your posting tool takes the MP4 from there.

Read the automation guide

Your scheduler6:00 wakes up
Your agent or scriptPicks today's approved topic
DoodaroPictures, narration, MP4
Third-party toolUploads or schedules the post

Safety & control

You stay in charge.

The agent asks first. The agent guide tells it to confirm with you before large generations or renders, and never to invent style or voice ids.
Keys you can revoke. Use one key per agent. Revoke it in Developer and the next request fails with 401.
Everything is visible. Every project the agent makes is a normal project in your dashboard. Open it, redraw one scene, render again.
The key stays secret. The agent guide tells the agent never to print the key. Keep it out of shared chats and code repos.

Beta

What the beta does not do yet.

Exported MP4s do not include captions yet, so turn on the platform's auto-captions when you upload.

Questions about AI agents and Doodaro

Can ChatGPT make videos with Doodaro?

Not yet. ChatGPT adds outside tools through a sign-in connector, and the Doodaro sign-in connector is coming soon. Today you can make Doodaro videos from Claude Code, Codex, Cursor or any MCP client that can send an API key header.

Can I use Doodaro with Claude?

Yes, with Claude Code. Add the Doodaro MCP endpoint with your API key, or paste the Copy for AI text. The Claude chat app needs a sign-in connector, which is coming soon.

What is MCP?

MCP (Model Context Protocol) is an open standard that lets an AI app call tools on another service. Doodaro MCP turns every Doodaro API action, such as create_project or render_video, into a tool your agent can call.

Do I need to code to make videos with an AI agent?

You do not write code, but you do need a coding agent such as Claude Code, Codex or Cursor installed, and you paste one command or one block of text. If that is not for you, the Doodaro dashboard does everything the agent does.

Can AI run my faceless YouTube channel?

An agent can take on most of the production: drafting scripts, building scenes, generating pictures and narration, and rendering. You still pick the niche, review every script and publish. YouTube asks creators to avoid repetitive, mass-produced content, so review every script and vary your videos. Nobody can promise views or income.

Does Doodaro post to YouTube for me?

Not yet. Doodaro Publish, which posts and schedules to YouTube and other platforms from the dashboard, is coming soon at no extra cost. Today Doodaro gives you an MP4 link. Upload it in YouTube Studio, or connect your own automation to the YouTube Data API or a scheduling tool such as Buffer, Metricool, Publer or Later. Those are third-party tools, so check their current plans and platform support.

Is it safe to give my AI agent a Doodaro API key?

Use a separate key for each agent and treat it like a password. You can revoke any key in Developer, and the next request with it fails. The agent guide tells the agent never to print the key back and to confirm with you before large generations or renders.

How much does it cost to make videos with an AI agent?

During the beta, minutes are not metered. Once metering is switched on, paid plans count minutes on rendered exports, from 120 minutes a month on Starter. Your AI agent is billed separately by its own provider.

More detail: Copy for AI and MCP setup · Developer hub · Automate daily videos

Rather try the studio first?

Everything your agent does, you can do by hand in the dashboard. Start with 2 free minutes, no card.

Start free See how it works