BrandGen v2 - On-brand AI creatives for marketers and founder
by•
Most AI image tools ignore your brand. BrandGen learns it first. Upload one reference or paste your guidelines, and it extracts your Brand DNA: colors, tone, composition. Then every image you generate stays on-brand.
v2 is a full workspace. A project per brand keeps your DNA loaded. Pick from 4 image models, enhance your prompt, generate up to 5 variants at once, then refine, share, and remix any result. On-brand marketing images in seconds, no design skills. Built for marketers and founders.
Replies
Hey Product Hunt! 👋 I'm Yash, founder of @brandgen_ai .
I've spent 9+ years shipping production AI, and one problem kept coming up with every marketing team I worked with: AI makes beautiful images that look nothing like your brand. Wrong colors, generic fonts, that "made by AI" look. So you fix every output by hand, or fight prompts for an hour and still ship off-brand.
So we rebuilt BrandGen from the ground up to fix exactly that.
🧠 The idea: learn your brand first. Upload one reference or paste your guidelines, and BrandGen extracts your Brand DNA - colors, typography, tone, composition. From there, every image you generate stays unmistakably you.
✨ What's new in v2:
A project per brand, so your Brand DNA stays loaded
4 image models to choose from
An inline editor to refine without tool-hopping
Batch variations in one shot
Share links
🍊 Launch-day gift: sign up today and get 100 bonus credits on top of the free 30. 24 hours only.
Free to start, no card. Paid plans from $7/mo when you need more.
I'd genuinely love your honest feedback: what makes an AI image feel "off-brand" to you? I'm here all day. 🙏
- Yash
How does BrandGen actually handle typography in the generated images — does it pull fonts from your uploaded assets or do you have to specify them separately for the brand DNA to stick consistently?
@birglakan4zsk Good question,
BrandGen reads typography cues from the uploaded asset, things like font style, weight, casing, hierarchy, spacing, and overall layout direction.
It does not magically “install” the exact font file from the image. If you want stricter typography consistency, the best workflow is to upload a reference and also mention the font/style in the guidelines. That gives the Brand DNA much stronger direction.
Typography is honestly one of the hardest parts of AI image generation, so we built BrandGen to treat it as part of the brand system, not just decoration.
Uploaded our style guide and the generated images actually matched our palette and tone surprisingly well, which is rare for these tools. Excited to see what the inline editor brings in v2.
@hkristinabeoigp Love hearing this 😁
Style guides were a big focus for v2 getting the palette, tone, and overall brand vibe right instead of just making something look "nice."
And yes, the inline editor is for that last 10%. small tweaks, copy fixes, and layout changes without starting over.
You can actually try v2 now it’s live today at getbrandgen.com
congrats for shipping @yashs369 qq. how well does the generator handle rendering clear, legible text inside the brand style when working with the Nano Banana Pro engine?
@vikramp7470 Thanks Vikram, good question.
Nano Banana Pro is much better with short, clear text than the older image models, especially when the copy is simple and the layout has enough space. That said, I would still not pretend text rendering is “solved” for every use case. Longer sentences, tiny labels, or dense typography can still need a refinement pass.
For BrandGen, the goal is to keep the text treatment inside the brand style, font feel, hierarchy, colors, spacing, while giving you room to refine if the exact wording needs cleanup.
What kind of text-heavy asset are you thinking of testing?
The color extraction actually nailed my brand palette on the first try, which I did not expect. Excited to see v2 with the editor and multiple models.
@tahsincumbul Thank you, that means a lot! Palette matching is one of the core things I really wanted BrandGen to get right, so hearing it nailed yours on the first try is amazing feedback. v2 is focused on giving you more control after generation, with the refine flow, and multiple model options. Would love to see what you created with it :)
Uploaded a moodboard and the generated images actually picked up on the muted palette and soft serif vibe i had going. Curious how much more control the inline editor will give in v2.
@arife364310 Thanks for actually testing it, glad the muted palette and soft serif came through, that's exactly what the brand extraction is built to catch.
The inline editor in v2 is a big step up: you can refine an image directly instead of regenerating from scratch, and each edit saves as a new version so you can compare. Curious what you'd most want to adjust after generating?
Hi PH! Abhishek here, I led engineering on v2.
The hard part of "on-brand" isn't generating one good image. It's the 50th still looking like you. We rebuilt the whole product around persistent brand projects so your Brand DNA carries across every generation, and added an inline editor so you refine without tool-hopping.
Months of work. Genuinely proud of this one. Try it and tell us where it nails your brand and where it misses 🙏
Hello... Tanmay here. I work across product and the v2 interface.
We wanted "upload your brand → on-brand asset" to feel like ten seconds, not a tutorial. Every screen is built so a non-designer never feels lost. If anything in the flow feels clunky, tell me, that's exactly the feedback I want today
Hey PH! Raj here 👋 — backend & infra on v2.
The invisible problem: "on-brand every time" isn't a generation challenge, it's an infra one. Persistent brand projects mean Brand DNA has to stay perfectly in sync across every generation and edit, at scale, without lag.
We rebuilt the pipeline to keep that fast and consistent no matter how hard you push it. Invisible work, but it's what makes the "on-brand" promise actually hold up.
Go stress-test it upload, generate, edit a ton 😁