Show HN:Draw Fast - Real-time AI image generation based on drawings in a canvas
drawfast.ioHi, this is Steve from tldraw. This project is a deployment of our open source draw fast repo, see here: https://github.com/tldraw/draw-fast
However, we did NOT publish this website! We have no relation it and can’t make any promises about it. The site is using an early access API endpoint from Fal that should be turned off soon. I would prefer this link be removed.
So is this basically a fork of https://github.com/SawyerHood/draw-a-ui ?
Almost completely freezes my browse/phone on Android when I try to change the prompt. I can basically type one letter and then have to kill the browser.
On the plus side it remembers the letter. But typing by killing & restarting the browser for every letter kinda ... sucks. :]
Awesome! This is exactly what I’ve been looking for. Is the source code available somewhere?
I don't know what model this specific application is based on, but you can get something similar wit StableDiffusion + ControlNet ie `control_v11p_sd15_scribble` in the ControlNet github repo.
I feel like I need a tutorial because I cannot get it to draw anything
I was confused too, but I'm stubborn, eventually figured out:
Clicking "Draw Fast" creates a box with the label "a city skyline"—drawing in that box will render an image to the right as you draw.
Double-clicking the box that "Draw Fast" produces allows you to edit the tag which says "city skyline" by default—allowing you to change what it renders as you sketch.
And can you copy and/or save the generated image?
I can't seem to find an explicit save or export option apart from File > Print. Using that, the Chrome print dialog, "More Settings" allows it to be opened in Preview or "Print Using System Dialog" where it can be saved to PDF.
Alternatively, I just screen captured the rendered image with MacOs "copy selected region to clipboard" command.