TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: Using Stable Diffusion to generate predictable illustrations

4 pointsby Ivan_Vabout 2 years ago
Hey HN,<p>We are Kirill and Ivan. We&#x27;ve been working together on various projects since 2014, and our latest one is something we&#x27;re really excited about.<p>Since October 2022 we&#x27;ve been working on NewArc.ai (<a href="https:&#x2F;&#x2F;newarc.ai" rel="nofollow">https:&#x2F;&#x2F;newarc.ai</a>), a tool that allows creating beautiful illustrations with ease.<p>When Stable Diffusion was released publicly in September 2020, we played around with it, trying to find the best way to use it in the work of designers and illustrators.<p>We realised that text-to-image generation in Stable Diffusion was great for inspiration, but not suitable for creating final images because of its creativity. That&#x27;s why we combined image-to-image generation, ControlNet and the ability to train custom models (Dream Booth, LORA) to create a tool that allows illustrators to define composition, colour, and style to generate images with predictable results.<p>Today we&#x27;re releasing the first version of our service for illustrators, and here’s how it works:<p>1. Upload a sketch to define the composition and colours. It will allow you to generate predictable draft illustrations.<p>2. Select one of our standard styles or add your very own style. Generate dozens of drafts that match your style, colour, and composition requirements.<p>3. Choose the best draft illustration that matches your vision, then add the finishing touches in your favourite graphics editor to make it perfect.<p>So illustrators can say goodbye to tedious work, save time and get more done or enjoy more freedom.<p>Happy to further discuss details in the comments!

no comments

no comments