Wireframing 1.0
Wireframing 2.0
Library coverage
Uses ~25% of marketing components.
Uses ~70% of marketing components.

More variety and creative layouts appear in generations.
Brand style fit
Mixed styles, not brand-aware.
Selects styles that match the project description (e.g., off-grid/overlapping for agencies; card layouts for SaaS).

Wireframes feel “on-brand” from the first draft.
Section content match
Defaulted to generic layouts (e.g., always 3 pricing plans), relying only on the section title.
Selects components based on content count using both title & description (e.g., “2 pricing plans” generates a 2-plan layout).

Sections communicate the right thing with the right structure.
Page flow awareness
No awareness of position/order.
Chooses components with placement in mind; improves rhythm and alternation.

Smoother narrative and fewer repetitive beats down the page.
Site-wide consistency
Components are not reused across pages.
Reuses suitable components across pages for coherence.

A more unified site with less rework later.
Copywriting 1.0
Copywriting 1.5
Context
Only looked at one section at a time.
Considers the entire page.

Less repetition, smoother flow across sections
Model
OpenAI GPT
Claude

Sharper, more natural copy.
Quality
Generic, often placeholder text.
More relevant, compelling, and usable.

Saves editing time, closer to production-ready.
Consistency
No awareness of position/order.
Unified tone across the whole page.

Feels like one voice, not stitched-together fragments.
Andy McKay
Principal Product Designer
Lorem ipsum
Sam Parkinson
Snr. Software Engineer
Lorem ipsum
Sanskar Gauchan
Software Engineer
Lorem ipsum
Team
Andy, Sam & Sanskar
Andy McKay
Principal Product Designer
Lorem ipsum
Sam Parkinson
Snr. Software Engineer
Lorem ipsum
Team
Andy & Sam
Damian Martelli
Lead Product Designer
Lorem ipsum
Sam Parkinson
Snr. Software Engineer
Lorem ipsum
Team
Damian & Sam
Damian Martelli
Lead Product Designer
Lorem ipsum
Sanskar Gauchan
Software Engineer
Lorem ipsum
Team
Damian & Sanskar
Damian Martelli
Lead Product Designer
Lorem ipsum
Meredith Lane
Snr. Software Engineer
Lorem ipsum
Team
Damian & Meredith
Damian Martelli
Lead Product Designer
Lorem ipsum
George Caley
Snr. Software Engineer
Lorem ipsum
Team
Damian & George
Damian Martelli
Lead Product Designer
Lorem ipsum
Chester Liu
Snr. Software Engineer
Lorem ipsum
Maria Marin
Product Designer
Team
Damian, Maria & Chester
Rob Austin
Snr. Software Engineer
Lorem ipsum
Nour Jandali
Front-end Developer
Lorem ipsum
Team
Rob & Nour
Kaleb Jae
Component Team Lead
Lorem ipsum
Maria Marin
Product Designer
Jonathan Nathan
Product Designer
Lorem ipsum
Team
Kaleb, Maria & Jonathan
Kaleb Jae
Component Team Lead
Lorem ipsum
Maria Marin
Product Designer
Team
Kaleb & Maria
Kaleb Jae
Component Team Lead
Lorem ipsum
Jonathan Nathan
Product Designer
Lorem ipsum
Damian Martelli
Lead Product Designer
Lorem ipsum
Team
Kaleb, Jonathan & Damian
Kaleb Jae
Component Team Lead
Lorem ipsum
Damian Martelli
Lead Product Designer
Lorem ipsum
Jonathan Nathan
Product Designer
Lorem ipsum
Maria Marin
Product Designer
Team
Kaleb, Damian, Jonathan & Maria
Damian Martelli
Lead Product Designer
Lorem ipsum
Maria Marin
Product Designer
Marc-Antoine Roy
Principal Product Designer
Lorem ipsum
Team
Damian, Maria & Marc
Damian Martelli
Lead Product Designer
Lorem ipsum
Maria Marin
Product Designer
Kaleb Jae
Super Mega Manager Guy 3000
Lorem ipsum
Team
Damian, Maria & Kaleb
Kaleb Jae
Component Team Lead
Lorem ipsum
Matt Johnson
Head of Support
Lorem ipsum
Team
Kaleb & Matt
Kaleb Jae
Component Team Lead
Lorem ipsum
Matt Johnson
Head of Support
Lorem ipsum
Jonathan Nathan
Product Designer
Lorem ipsum
Team
Kaleb, Matt & Jonathan
Andy McKay
Principal Product Designer
Lorem ipsum
Chris Morgan
Snr. Software Engineer
Lorem ipsum
George Caley
Snr. Software Engineer
Lorem ipsum
Kaleb Jae
Guy with ideas
Lorem ipsum
Damian Martelli
Lead Product Designer
Lorem ipsum
Maria Marin
Product Designer
Lorem ipsum
Team
Andy, Chris, George, Kaleb, Damian & Maria
Damian Martelli
Lead Product Designer
Lorem ipsum
Trung Luu
Senior Software Engineer
Lorem ipsum
Andy McKay
Principal Product Designer
Lorem ipsum
Weidong Fang
Senior Software Engineer
Lorem ipsum
Kaleb Jae
Super Mega Manager Guy 3000
Lorem ipsum
Maria Marin
Product Designer
Lorem ipsum
Team
Damian, Trung, Andy, Weidong, Kaleb & Maria
Nick Flower
Growth Product Design Lead
Lorem ipsum
Abhi Lamba
Head of Growth
Lorem ipsum
Team
Nick & Abhi
Nick Flower
Growth Product Design Lead
Lorem ipsum
Weidong Fang
Senior Software Engineer
Lorem ipsum
Team
Nick & Weidong
Nick Flower
Growth Product Design Lead
Lorem ipsum
Weidong Fang
Senior Software Engineer
Lorem ipsum
Maksim Salomatin
Snr. Software Engineer
Lorem ipsum
Team
Nick & Max
Andy McKay
Principal Product Designer
Lorem ipsum
Weidong Fang
Senior Software Engineer
Lorem ipsum
Team
Andy & Weidong
Andy McKay
Principal Product Designer
Lorem ipsum
Andy McKay
Principal Product Designer
Lorem ipsum
Trung Luu
Senior Software Engineer
Lorem ipsum
Team
Andy, Diego & Trung
Andy McKay
Principal Product Designer
Lorem ipsum
Trung Luu
Senior Software Engineer
Lorem ipsum
Weidong Fang
Senior Software Engineer
Lorem ipsum
Team
Andy, Trung & Weidong
Marc-Antoine Roy
Principal Product Designer
Lorem ipsum
Trung Luu
Senior Software Engineer
Lorem ipsum
Team
Marc & Trung
Marc-Antoine Roy
Principal Product Designer
Lorem ipsum
Arman Mohajer
Senior Software Engineer
Lorem ipsum
Team
Marc & Arman
Marc-Antoine Roy
Principal Product Designer
Lorem ipsum
Keegan Cruickshank
Snr. Software Engineer
Lorem ipsum
Team
Marc & Keegan
Marc-Antoine Roy
Principal Product Designer
Lorem ipsum
Shaun Donnell
Snr. Software Engineer
Lorem ipsum
Team
Marc & Shaun
Abhi Lamba
Head of Growth
Lorem ipsum
Team
Pantelis

Introducing Relume Studio

Today we’re launching Relume Studio, so you can generate on-brand images with the same look, the same people and the same places every time. Lock them in once, and every image after stays true, with no photographer, location or half day of shooting.

Realistically, AI image generation usually means one great picture and then a lot of drift. Ask for a second image and the face is slightly off. By the fifth it’s a different person, the office has moved and the look has slid into that generic AI aesthetic. With Relume Studio, your style, characters and scenes are locked from the start, so every image references the source rather than the last thing it made.

In the demo video below, Andy walks through how we used Studio to launch Studio, from pasting in relume.io to dropping the finished images into our launch page. If you’d rather skim, we’ve summarised what Studio is, why we built it and how it works below.

What is Relume Studio?

Relume Studio is where your brand’s imagery lives. It’s built on three building blocks: Styles lock the look, Characters lock the people and Scenes lock the place.

Each one can be built by hand, exactly how you want it, or BrandKits can build all three for you from a single URL. Select them in the composer and every image you generate respects all three at once.

Everything you make is saved to your library, ready to download or drop straight into a Relume site.

Why we built it

Most businesses can’t afford a photographer every time they need a new image, and stock photos never look like the actual business. AI image tools looked like the answer, but for a brand they fall over fast. You can’t run a campaign on images where the founder looks different in every shot.

We wanted Studio to solve the consistency problem first. If the face, the space and the look don’t hold, nothing else matters. So rather than building another prompt box, we built a way to lock things in.

It also needed to fit into how you already build. The images you make shouldn’t end up in a downloads folder. They should be ready to use on the site you’re working on.

‍

Styles lock the look

Start with references you love and drop them in. Studio creates new seeded references from them, so you get a locked look rather than a copy of someone else’s image.

A style locks more than colour. It locks how close the camera sits, whether people look at the lens or away, and whether the result feels polished or cheesy. There are four kinds: photography, illustration, graphics and mocks.

Styles can also be guided. Add a rule like “take the colours from whatever I attach” and the style adapts to each asset you give it. Mocks keep a scene locked and place your content into it. In the demo, our launch graphic lands on a billboard in a Sydney train station.

Characters lock the people

Upload a few photos and Studio builds a full reference sheet, with every angle and every expression. Every new image goes back to that sheet rather than the previous image, so the face never drifts.

That makes it a good fit for founders who are the face of their brand, or teams who want real staff in their marketing rather than stock photos.

Scenes lock the place

Same idea, but for places: an office, a gym, a garage or a shopfront. One photo gives you four angles of the space, and Studio locks where things actually sit, so the desk, the window and the shelf stay put. Add more photos and the lock gets tighter.

Put it all together and you get you, in your space, in your brand’s style. My face, my space, our brand. Every time.

Straight into your site

Everything you make in Studio is already in the library inside Publish. Open a page, open the library and drop your images in. There’s no exporting, uploading or renaming files.

Share your styles

Styles can be shared publicly, so you can use styles other people have built or share your own.

Get started

Relume Studio is available now in beta on all Pro plans. Head to relume.ai/studio, paste in your URL and see what Studio builds for you.