# GetLayers Evaluation

## Verdict

GetLayers is good at art direction and weak as a production guarantee.

The two free prompts produced visually distinct desktop work with real signature interactions. They did not produce generic card-grid sites. The prompts are unusually detailed, with exact assets, colors, layout values, animation timings, and interaction logic.

They still need a competent developer. Mobile composition, contrast, performance, accessibility, real content, and production integration remain your problem.

**Recommendation:** do not buy Unlimited for $139 as a prompt-only impulse purchase. If a real client project can use this style immediately, buy Full Stack at the current $199 lifetime price. The extra $60 adds source files, MCP access, 3D scenes, backgrounds, live tuning, and support. That is the part that reduces generation variance and rework.

## What was tested

- Two complete free GetLayers prompts were copied from the public free library after account creation.
- Both were handed to the same coding model and implemented as standalone HTML, CSS, and JavaScript.
- No paid GetLayers source files or paid prompts were used.
- Prompt defects were left visible instead of silently redesigning the pages.
- Browser QA covered 1,440 x 1,000 and 390 x 844 viewports.
- Interactions tested: loader, navigation overlay, liquid pointer reveal, modal form success, scroll sections, 3D loading, pointer parallax, bubbles, and flavor switching.

## Results

| Category | Lumora | Soda |
| --- | ---: | ---: |
| Desktop visual quality | 8.5/10 | 9/10 |
| Mobile composition | 6/10 | 3/10 |
| Prompt specificity | 9/10 | 9/10 |
| Performance posture | 6/10 | 3/10 |
| Production readiness | 6/10 | 3/10 |
| Distinctiveness | 8/10 | 9/10 |

### Lumora

What worked:

- Strong editorial composition with a real signature interaction.
- Full page hierarchy holds together beyond the hero.
- Loader, navigation overlay, contact modal, success state, section reveals, and count-up behavior work.
- No horizontal overflow at either tested viewport.
- No browser errors in the QA run.
- Local synthetic metrics: 400 ms FCP, 1,560 ms LCP, and about 0.004 CLS.

What failed or needs work:

- The black mobile headline sits on the darkest part of the portrait and becomes difficult to read.
- The prompt specifies a dramatic hero but does not provide a robust contrast strategy for every crop.
- The two hero images total about 5.2 MB before fonts and JavaScript.
- Portfolio content is visually repetitive because the prompt provides the system and copy, not four rich project art directions.
- The build is a strong starting point, not a shippable client site.

Evidence:

- [Desktop hero](evidence/lumora-hero-desktop.png)
- [Desktop full page](evidence/lumora-full-desktop.png)
- [Mobile hero](evidence/lumora-hero-mobile.png)
- [Mobile full page](evidence/lumora-full-mobile.png)
- [Mobile menu](evidence/lumora-menu-mobile.png)
- [Form success](evidence/lumora-modal-success.png)

### Soda

What worked:

- This is the stronger proof that GetLayers can escape generic AI design.
- The 3D can, floating fruit, bubbles, glass navigation, and flavor choreography create a coherent scene.
- Fourteen visible model-viewer elements loaded successfully in the completed QA state.
- Classic to blue flavor switching updates the theme, active card, can texture, and fruit models.
- The page produced no runtime errors during the completed desktop and mobile passes.

What failed or needs work:

- On a 390 px viewport, the product carousel extends to about 431 px and is clipped by the no-scroll body.
- The right headline is also visibly cropped on mobile.
- One cold mobile run exceeded the 25 second ready-state timeout. A cached rerun completed within 12 seconds.
- Unique GetLayers scene assets total about 8.64 MB before GSAP, model-viewer, fonts, and other overhead.
- The page creates 16 model-viewer elements, including two preload instances. This is spectacle with a real device cost.
- The prompt's responsive section contains only a broad 1,200 px breakpoint and does not solve phone composition.

Evidence:

- [Classic desktop](evidence/soda-classic-desktop.png)
- [Blue desktop](evidence/soda-blue-desktop.png)
- [Classic mobile](evidence/soda-classic-mobile.png)
- [Blue mobile](evidence/soda-blue-mobile.png)

## Product and pricing read

The public pricing page says:

- Unlimited: $139 lifetime for the complete prompt library and commercial license.
- Full Stack: currently shown as $199 lifetime, reduced from $759.
- Full Stack adds source code and the pieces that require real files or live canvases.
- MCP is included with Full Stack Lifetime.
- Purchases are not refunded mid-cycle because the digital content unlocks immediately.

The site's own documentation also says some templates still need mobile performance and load-speed optimization. That admission matches this test.

Sources:

- [GetLayers pricing](https://www.getlayers.ai/pricing)
- [GetLayers documentation](https://www.getlayers.ai/docs)
- [GetLayers MCP](https://www.getlayers.ai/mcp)

## Ruthless buying rule

Buy Full Stack only if all three are true:

1. You have a real site or client build that can use one of these visual systems in the next 30 days.
2. One saved design day or one client payment covers the $199.
3. You accept that mobile QA and performance cleanup still require engineering time.

Otherwise, keep the $199. The free prompts already prove the concept, and another prompt library will not fix a weak offer, weak copy, or weak execution.