Google Pomelli Review: 7 Things It Does Well (and 2 Limitations to Know Before You Start)
Pomelli generates 3 campaign concepts from one product image and a target audience prompt. Here's what works, what doesn't, and the animation text bug to avoid.
Google Pomelli Just Launched, and the Campaign Generator Is the Part Worth Paying Attention To
Google Pomelli is available now through Google Labs, and the headline feature is this: drop in a single product image, describe your target audience, and get 3 campaign concepts generated in seconds. That’s the pitch. Seven things about it work well enough to matter. Two things will trip you up if you don’t know about them going in.
This isn’t a tool for enterprise marketing teams with dedicated brand studios. It’s aimed squarely at small business owners, solo operators, and anyone who needs to go from “I have a product” to “I have social-ready creative” without a design budget. Whether it actually delivers on that depends a lot on which features you lean on — and which ones you avoid.
Here’s the honest breakdown.
The Campaign Generator Is the Core, and It’s Genuinely Useful
The workflow is simpler than it sounds. You upload one product image — ideally the product on a plain background, nothing elaborate — describe who you’re targeting, and Pomelli generates three distinct campaign concepts. Each concept comes with a headline, a description, and a visual treatment that places your product in context.
Remy doesn't build the plumbing. It inherits it.
Other agents wire up auth, databases, models, and integrations from scratch every time you ask them to build something.
Remy ships with all of it from MindStudio — so every cycle goes into the app you actually want.
The key word there is context. Pomelli isn’t just slapping your product image onto a template. It’s extracting the product from the image and resituating it — a cookie on a plain white background becomes a cookie in a nighttime street scene, or held by someone at a party, or arranged with ingredients. The campaign concept shapes the visual, not just the copy.
In the demo walkthrough, a cookie product uploaded with the prompt “target individuals on a night out” produced three concepts: “Status Unlocked” (freshly baked cookies delivered across London until the early hours), “Drop the Box” (reveal the cookies, watch the mood shift), and a third variation. These aren’t generic. They’re specific to the audience framing you gave it.
The practical implication: the quality of your target audience description directly affects the quality of the output. “Target late-night gamers” produces different concepts than “target individuals on a night out.” You’re essentially writing a brief, and the tool responds to brief quality.
The Fix Layout Button Solves a Real Problem
One of the more quietly useful features is the Fix Layout button inside the image editor. When a generated creative comes back with the product in an awkward position — too small, pushed to an edge, compositionally off — you don’t have to regenerate from scratch. You click Fix Layout, and Pomelli analyzes the image and reorients the product placement automatically.
This matters more than it sounds. Regenerating an image from scratch means losing the specific visual treatment you liked. Fix Layout preserves the concept while correcting the composition. In practice, it’s the difference between a tool that requires you to get lucky on the first try and one that lets you iterate toward something usable.
The editing layer more broadly is more capable than the campaign generator’s surface suggests. You can toggle the header and description on or off, update copy directly, change font, adjust color, resize text, and control the call-to-action. The left/right arrows let you cycle between versions, so you can compare the original against your edits without losing either.
Business DNA Is What Makes the Whole Thing Coherent
Pomelli’s Business DNA system is the connective tissue. It stores your brand values, brand aesthetic, brand tone of voice, business overview, logo, colors, and fonts — and every piece of creative generated pulls from it. Without it, you’re just generating generic AI images. With it, the outputs stay on-brand across campaigns.
The setup process has a useful shortcut that most people will miss. When you start from a template (say, “food and beverage brand”), Pomelli gives you pre-filled values to work from. The recommended move: screenshot those template fields, paste the screenshot into Gemini with a prompt like “I’m starting a new cookie shop business, help me come up with the values to plug in here,” and let Gemini generate three or four concept directions with fully fleshed-out values for each field. You’re not brainstorming from zero — you’re selecting from AI-generated options.
The same trick works if you already have a brand. Take a photo of your packaging, paste it into Gemini, and ask it to extract brand values, aesthetic, tone of voice, and business overview from the image. The output maps directly to Pomelli’s fields.
Logo Generation Works, With Caveats
Other agents start typing. Remy starts asking.
Scoping, trade-offs, edge cases — the real work. Before a line of code.
Pomelli can generate a logo from a text prompt. You describe what you want, hit enter, and get four concepts. You can regenerate, cycle through previous attempts, and apply the one you want directly to your Business DNA.
The practical workflow here is to use AI to write the prompt rather than writing it yourself. In the demo, the approach was: describe your brand concept to an AI, ask it to generate three logo description prompts without rendering any images, then paste those prompts into Pomelli one at a time. This produces more specific, on-brand results than “make me a logo for a cookie company.”
The color picker follows from the logo. The recommended approach is to pick your two main colors directly from the logo you generated, then optionally add intermediate colors. Pomelli’s color system is intentionally simple — it’s not a full brand color management tool — but it’s enough to keep your campaign outputs visually consistent.
For fonts, the goal is to match the aesthetic of your logo. Pomelli includes options like Permanent Marker and Tilt Neon, among others. If your logo has a handwritten feel, Permanent Marker. If it’s more electric and modern, Tilt Neon. The font choice propagates through your campaign creatives, so it’s worth spending a few minutes here rather than accepting the default.
The Photoshoot Feature Generates Four Distinct Asset Types From One Image
The Photoshoot feature is separate from the campaign generator, and it solves a different problem. Upload a product image, and Pomelli generates four professional-style photos across four templates: Studio (clean product shot), Ingredient (product surrounded by its components), In Use (someone using or consuming the product), and Contextual (product in an environment where it would naturally appear).
Pomelli auto-selects the relevant templates based on product type. A food product won’t get the beauty template. A skincare product won’t get the ingredient template in the same way a food product would. The auto-selection is smart enough that you can often just accept the defaults and click generate.
The editing capability here is worth highlighting. After generation, you can click into any image and change the background via a text prompt. “Change the background to pink” works exactly as described — the background changes, the product stays. This is the kind of iterative editing that makes a photoshoot tool actually useful rather than a one-shot generator.
When you’re done, there’s a “Download all assets” button, and a one-click “Add all to Business DNA” option that saves your generated images to your brand library. The next time you run a campaign, those images are available as source material.
Animation Exists, But It Has One Hard Constraint
Pomelli can animate your campaign images into short videos. The output is a moving version of your creative — product in motion, text appearing, scene animating. For simple products, it works well enough to be usable.
There are two things to know before you try it.
First, you have to change your image to 9:16 story format before the Animate button appears at all. It’s not available on square or landscape crops. This isn’t documented prominently, and it will confuse you the first time.
Built like a system. Not vibe-coded.
Remy manages the project — every layer architected, not stitched together at the last second.
Second — and this is the more important one — animated text in videos “usually gets messed up,” as the source walkthrough puts it directly. The recommendation is to animate without text, then add text as a separate layer in your video editor if you need it. This is a current limitation, not a design choice, and Google will likely address it. But right now, if you animate with text on, expect distorted or broken lettering in the output.
The animation itself doesn’t take prompts. You click Animate and Pomelli decides how to move the scene. You can regenerate for variations, but you can’t direct the motion. For a product sitting on a surface, this is usually fine. For anything with complex action, it’s limiting.
The Gemini Integration Is Underused by Most People
Pomelli is a Google Labs product, which means Gemini is available as a companion throughout the workflow. Most people treat it as an optional add-on. It’s actually the most efficient part of the setup process.
The screenshot-to-Gemini workflow for Business DNA setup is the clearest example. But it extends further: if you’re stuck on what your target audience description should be, ask Gemini. If you need a logo prompt, ask Gemini. If you want to generate three variations of your brand tone of voice before committing to one, ask Gemini. The tool is sitting right there, and using it as a thinking partner rather than a fallback makes the whole Pomelli workflow faster.
This kind of multi-model orchestration — using one AI to set up inputs for another — is increasingly how production marketing workflows get built. Platforms like MindStudio handle this at scale: 200+ models, 1,000+ integrations, and a visual builder for chaining agents and workflows, which is useful when you’re moving beyond one-off campaign generation into repeatable automated pipelines.
Limitation 1: Animation Text Is Broken Right Now
This deserves its own section because it’s the kind of thing that will waste your time if you don’t know about it. The animation feature generates video from your campaign images, and it looks good when it works. But if your image has text overlaid — a headline, a description, a call-to-action — that text will almost certainly render incorrectly in the animated output. Distorted, repositioned, or illegible.
The workaround is to strip the text before animating. Generate your image with text for the static version, then remove the text layer, animate the clean image, and composite the text back in post if you need it. It’s an extra step, but it’s the only reliable path right now.
This is a known issue, not an edge case. Plan around it.
Limitation 2: No Prompt Control Over Animation Direction
The second limitation is less of a bug and more of a design gap. When you animate an image, Pomelli makes all the decisions about how the scene moves. There’s no prompt field, no direction controls, no way to specify “zoom in on the product” or “pan left” or “text enters from the bottom.”
For simple products in simple scenes, the auto-animation is usually acceptable. For anything where the motion matters — a product demo, a before/after reveal, a specific brand moment — you’re at the mercy of what Pomelli decides. You can regenerate for variations, but you can’t steer.
Remy is new. The platform isn't.
Remy is the latest expression of years of platform work. Not a hastily wrapped LLM.
This is where Pomelli sits clearly in the “good enough for social” category rather than “production video” territory. If you need directed animation, you’ll need to export the static image and use a dedicated video tool. The AI agents for marketing teams use case is real, but animation direction is still a gap in what single-tool solutions can handle end-to-end.
So What Does Pomelli Actually Add Up To?
Pomelli is a focused tool. It does one thing well: take a product, a brand identity, and a target audience, and generate social-ready creative without requiring design skills or a large team. The campaign generator, the photoshoot templates, the Business DNA system — these work together in a way that feels considered rather than bolted together.
The comparison worth making is to what it replaces. If you’re currently hiring a freelance designer for every campaign iteration, or spending hours in Canva trying to make things look on-brand, Pomelli compresses that to minutes. The output quality isn’t at the level of a skilled human designer. But the speed-to-usable-asset ratio is genuinely different from what existed before.
The animation limitation is real and will frustrate people who don’t know about it. The lack of prompt control over motion is a gap that matters for certain use cases. And the tool is still in Google Labs, which means the feature set is actively changing — the animation controls in particular seem like an obvious next addition.
For AI builders thinking about where tools like this fit in a broader stack: Pomelli is a good example of a vertical AI application that handles a specific workflow end-to-end. The interesting design question is what happens when you want to extend it — connect it to a CMS, trigger campaigns based on inventory data, or chain it with other models. That’s where spec-driven tools like Remy become relevant: you write the integration logic as annotated markdown, and the full-stack application — backend, database, auth — gets compiled from it, rather than hand-coding every connection.
The image generation space is moving fast. GPT Image 2 handles product packaging and UGC ads with strong text rendering. Google Stitch can extract a design system from any URL and export it to React. Pomelli sits in a different lane — it’s not a general-purpose image generator, it’s a campaign workflow tool — but it’s competing for the same budget and attention from small business owners who need creative output without creative overhead.
The verdict: use it for campaign ideation and product photoshoots. Avoid animating with text until Google fixes it. And spend the extra ten minutes on Business DNA setup — it’s the part that makes everything else work.
