Brandmark
p/brandmark
The deep learning logo maker
Jack Qiao
Logo Rank — Let an AI critique your logo design
Featured
41

A neural net that rates your logo design on 3 main factors: uniqueness, legibility, and color/contrast

Replies
Chad Whitaker
Still going strong 😎
Sourav Ray
I think it's logo for my next project doing pretty good (considering my poor designing literacy)
David Yo
Uniqueness: 96 Legibility: 99 Color: 59 (it's white on black) Overall: 85
Ghost Kitty
Comment Deleted
Zach R
https://cl.ly/3I0i471R3I1Z - The pickle gets a 99. lol
Jack Qiao
@bluetidepro pickle rick! but actually photos could throw it off since it's trained on icons.
Ankit Dubey
I am not well acquainted with design or logo. Few days back i did tried making icons/logo stuff on different services, my experience with most of them were bad either templates sucked or logo making experience was in disarray or clues to what i was doing was good or bad were missing. I gave this a shot. Logo rank/ brandmark ...feels awesome. Logo making experience was smooth, didn't feel confused. Happy to use. Awesome work guys.
Leandro
Cool concept, well executed. Got a bit carried away checking out various logos!
zayle
Great idea for my logos. :D
Alexandre Vallières-Lagacé
Uniqueness: 99 Legibility: 86 Color/Contrast: 100 Overall: 95 Pretty proud for a non-designer :P http://cl.ly/mBVq
Dhaval Chauhan
Let's see who is hotter :p
Lance Cameron
http://prntscr.com/gd0928 I got 99 percent - cool tool!
Jack Qiao
hey guys you might have seen my other project https://www.producthunt.com/post... this is the same neural net that I use to evaluate icons for use in logos. it rates an icon's suitability for logos based on 3 categories: - uniqueness: the neural net compares the icon to all the icons on thenounproject (a bit more than a million), and lets you know if it looks like a generic icon. - legibility: the neural net scores the input image based on how legible it looks. Very complex, illustrative images receive lower scores. Bold shapes with lots of low-frequency variation receive higher scores. - color and contrast: higher scores for use of color as well as image contrast. This one doesn't rely on a neural net but pixel-wise contrast and saturation. I added today's top products to the random list, skipflag scored the highest!
Stuart Campbell
@jak_9994 thanks. I threw up a Graphic Springs logo and got a 100 on uniqueness (it's one of their new ones and not greatly used, so I say this just as an FYI). We're remaking it before we launch but given the results, may not deviate too much.
Chris Messina
Top Hunter
@jak_9994 I wonder if @peteskomoroch is aware of that!
Andreas Adamides
Compared two very different logos. One got 53% and the other 96%. Worked brilliantly.
Jose Garay
This was the AI threat Elon talked about 😱😂 great work!
Samar Mustafa
Teamstack.co got 62% Overall.. need to rework on the logo perphaps. Legibility: 86 Color: 100 Overall : 62
Vincent Soo
I think the size matter if I upload a rectangle size it will crop and fail to show the uniqueness number, would be great if you could state there the ideal size to upload. Overall a very great product, enjoy it much!
Hammad Nasir
My startup's logo overall score is more than Apple's logo overall score. Repairer (my startup) - 80. Apple - 60. HEHE!
Joshua Pinter
@hammadnasir Apple got 60?! That makes no sense. Apple's logo is one of the best in the world. Simple, unique, high contracts, beautiful.
Jack Qiao
@hammadnasir @joshuapinter Basically that's because there are a lot of stock apple icons. You can build a brand from very generic looking symbols too - eg. the red cross
Jason Bell
I didn't expect to be so impressed when I came across this. Excellent implementation.
Miguel H Gonzalez

While not definitive, measurement and assessment of a design artifact can be supplemented by an impartial algorithm. The score can help with a subjective, human decision.

Pros:

Provides an impartial way to assess a graphic design project

Cons:

Human, subjective judgement is impossible to duplicate with an algorithm