UmZobo — AI image and video studio

Create with AI Where imagination takes form.

50+ models in one studio. Run the same prompt across GPT, Flux, Midjourney, Seedream, Nano Banana, Kling, Wan and Grok. A clip is up to 30 seconds — continue from the last frame and a flow runs to about 5 minutes.

Open Image Lab Generate video

50+ AI modelsSame prompt, many modelsFlows up to 5 minMulti-step flowsRef2V identity

The AI studio

Four engines, one prompt

Text → image · 50+ models

Same prompt. Every model. Best frame.

Photoreal, anime, sci-fi, fashion. Run one prompt across GPT, Flux, Midjourney, Seedream, Nano Banana, Grok and more — then keep the winner.

Open Image Lab

Text → image

Night salon

Hung like a piece of art

Stills from Image Lab. Fifty-plus AI models, one prompt each, no retouch. Tap a frame to open it.

by akshat 44 1

AI stories

Long reads, illustrated by the models.

Serialized fiction with AI stills on every chapter — read as a magazine, or switch to the visual-story skin.

The Last Key

A London locksmith is hired to open a sealed riverside vault. Inside is a dead journalist, a USB drive, and a brass key that matches the lock on the house she grew up in. Someone has been staging her life for twenty years.

thrillerenglish

Read Visual story

The whole AI studio, inside your agent.

An MCP server for Image Lab, video, edit, upscale and stories. Works with Cursor, Claude Code and any MCP client — 50+ models, same-prompt runs, long video with multi-step flows.

CursorClaude CodeAny MCP client

50+ AI modelsEvery Image Lab and video engine you see on the web, callable as tools.

Same prompt, many modelsTry GPT, Flux, Midjourney, Seedream and more on one prompt and keep the best.

Long video, multi-stepA clip is up to 30 seconds. Continue from the last frame and stitch a film up to about 5 minutes.

StoriesWrite, insert AI media, publish — from a prompt in your editor.

API-key authOne key in your MCP env. Billed at 2× web token rates.

Set up Connect+

install

# one line, then add your key to the MCP env
$ curl -fsSL https://ai.umzobo.studio/downloads/connect-plus/install.sh | bash

# tools you get
connect_generate_image · connect_generate_video
connect_edit_image · connect_upscale_image
connect_story_write · connect_story_insert_media

Tokens

Pay for AI frames, not seats.

No subscription. Buy a pack, spend it across 50+ image models, video, edits and stories. Tokens never expire.

Starter400tokens$15

~25 AI image generations
Edits & upscales
Works with Connect+

Get 400

Most popularPopular1,000tokens$30

~60 AI image generations
Short AI video clips
Priority rendering

Get 1,000

Studio5,000tokens$100

Studio-scale volume
Best per-token rate
Everything in Popular

Get 5,000

See all packs

Questions

What is UmZobo?

UmZobo is an AI image and video studio. Image Lab, the video studio, illustrated Stories, and the Connect+ MCP server are the product. You pay with token packs, not a subscription.

Which AI models does UmZobo support?

Image families in the catalog include GPT Image, Flux, Midjourney, Seedream, Nano Banana, Qwen, Wan, Grok, Hunyuan, and Z-Image. Video families include Wan, Kling, Grok, Veo, Seedance, Vidu, and Flux 3.

How long can AI videos be on UmZobo?

Wan 3.0 text-to-video is listed at 2–30 seconds per clip. Continue from the last frame and a flow can run to about 5 minutes.

How does UmZobo pricing work?

Token packs are 125 tokens for $5, 400 tokens for $15, 1000 tokens for $30, 2000 tokens for $50, 5000 tokens for $100. Tokens pay for image, video, edit, upscale, and stories.

UmZobo AI

Start with a sentence.Leave with AI art

Open Image Lab, pick a model — or run the same prompt across many — and make something worth framing

Start creating