Draw ThingsUnclaimed profile

On-device AI image and video generation for Mac and iOS — offline, private, with ControlNet, inpainting, and LoRA training.

Task: Generating Images & DesignUse case: Ai assisted image generation image to video and style transformation running loc

0.0

based on 0 reviews

Claim this Agent

Are you the vendor?

Be the first to ask to review this Agent

We'll let the team know there's interest. Counts are anonymous — no individual users are named on this page. This profile is not yet operated by a verified vendor.

reviews

Building the picture

Manzia generates an at-a-glance sentiment summary once an Agent has 5 approved reviews. This Agent has 0.

Video tutorials

Community tutorials

Summarized from 3 community tutorials · July 2026

What you need

Access: Free tier
Runs on: Desktop app
You’ll need: Mac (Apple Silicon recommended)
Mac App Store account
Downloaded model file (6–24 GB)
Skill level: Some setup — keys, settings, or a connected account

Worth knowing: Model files are 6-24 GB and slow to download; face-guided "control" generation from a reference photo was shown failing repeatedly to match the presenter's likeness.

Summarized from 3 community tutorials · July 2026

Highlights

Fully local, offline image generationGenerates images entirely on-device (Mac) with no internet, account login, or content filters required.
Multiple swappable AI modelsDownload and switch between models like Flux, Qwen Image, Z-Image Turbo, and Ernie Image Turbo from a built-in model browser.
Image-to-image inpainting repairsErase a masked region (e.g. garbled text or malformed hands) and regenerate just that area with a short prompt to seamlessly fix it.
Batch generation via JavaScript scriptsWrite or paste JS scripts that mix variables (titles, styles, materials) to auto-generate hundreds of image variations unattended.
Reference-image guided (control) generationLoad a custom image as a control input to steer composition/pose of new generations, with adjustable control weight and step range.

Summarized from 3 community tutorials · July 2026