There is no “best XenForo theme” in 2026. There is a Resource Manager snapshot, a free baseline you already own, and a short list of jobs a paid parent can still win: a framework you will child for years, a one-shot skin for a niche, or a gaming look you do not want to paint yourself. Anyone selling you a single winner is selling you their affiliate link.
This is a marketplace comparison. It is not a ranking. It is not the 2.3 variation switch — that article already owns Color palette Light/Dark, the footer gadget, and why you should not install a third-party “dark mode switcher” on 2.3. It is not the later “customize without breaking upgrades” process (child style, TMS, extra.less). It is not a light-versus-dark design essay. Stay here for what is actually on the Styles [2.x] category as of this writing, what those products cost when the RM lists a price, and how to compare them without a trophy.
The official manual’s pointer is still the right first URL: pre-built styles in the Resource Manager. Dates and prices below are from that marketplace snapshot (checked against the August 2026 RM listing used for this series). If a vendor has moved a number since you opened this page, believe the RM, not a blog.
What a XenForo style actually is
Official definition, because buyers still talk as if a “theme” were a wallpaper:
A Style is HTML templates + CSS/LESS templates + style properties. A new style with no parent inherits the hidden Master style. A child inherits its parent unless you override a property or a template.
Appearance options set the Default style for visitors who have not chosen, and a separate default style for HTML email. The style named “Default style” is just the one that shipped. You can rename it. Email can stay readable while the public variation is dark.
Properties inherit until customized. Types: colours, text, numbers with units, switches, multiple-choice, and CSS groups (text / background / border / padding / extra CSS). Color palette holds the chips. Basic colors assign chips to roles (@xf-paletteNeutral3, @xf-textColor). The Light/Dark style type switch tells intensify / mix / diminish which way to push a colour. Get the type backwards and every hover goes the wrong way.
Do not edit Master. Do not edit the vendor parent. Customizations live in a child. That sentence is the whole upgrade strategy. Direct template edits on the parent become outdated; after an upgrade you merge in Appearance → Outdated templates. Blue is auto-merged. Yellow is a conflict. Template modifications (Appearance → Template modifications) find/replace in the core template so the core stays uncustomized. They can bind to an add-on ID.
extra.less is not in the official styles manual. It is community standard: put custom CSS in the child’s extra.less; core upgrades do not overwrite it. Include extra files with <xf:include template="your_custom_file.less" /> (community). TMS into extra.less when several colour children should share a tweak. Treat that as practice, not as a documented API.
Custom font file upload steps are not in the official docs. Typography sizes and families exist as style properties. @font-face in extra.less is community practice. Self-host. Do not load Google Fonts from their CDN — site policy and performance, the same reason this blog does not.
Designer mode ($config['designer']['enabled'] = true) exists. It is advanced. You do not need it to pick a parent from the RM.
The free baseline: stock 2.3 plus variations
Before you spend anything, write down whether stock already does the job.
XenForo 2.3 style variations give you two slots on one style: default and alternate. They can follow prefers-color-scheme. Guests get a footer gadget. Preference is stored in cookies; 2.3.8 preserves that preference on logout (notebook). HSL CSS variables and xf-intensify / xf-diminish are variation-aware. The switcher article is the walkthrough. The 2.3 upgrade article is why you test both halves after you update.
That is a light and a dark of your Default style, not a third-party theme. For a lot of communities — support boards, internal forums, anything that should look like “a clean XenForo” — the honest move is:
- Appearance → Styles → Add style, child of Default.
- Enable variations on the child.
- Set palette chips per variation. Style type Light on the light half, Dark on the dark half.
- Change logo, header, and a handful of CSS-type properties.
- Stop.
You have not paid a vendor. You have not inherited a second outdated-template queue. You have not installed a framework add-on. You can still buy a parent later and re-parent the child if you accept the merge.
Buy a paid style when stock-plus-a-child cannot do the job: a magazine homepage chrome the portal tools will not fake, a gaming skin with a vertical rhythm you do not want to design, a framework your designer already knows. Do not buy a paid style because the Default header bored you on a Sunday. That is a property, not a product.
How to compare (the columns that matter)
Ignore “modern,” “stunning,” and demo databases full of fake threads. Compare these.
| Column | What you are asking | Why it bites |
|---|---|---|
| Framework vs one-shot | Will I child this for five years, or is it a finished skin? | Frameworks (UI.X 2, XenBase) expect an add-on and a workflow. One-shots look done until the next XF point release. |
| Update cadence | When did it last ship a build for this 2.3? | A beautiful 2.2 leftover is a merge job. Dates below are the snapshot. |
| 2.3 variations | Does it use the stock default/alternate slots, or a private switcher? | A private switcher is two styles again. You bought a 2.2 problem. |
| Add-on tax | What must be installed for the style to boot? | UI.X 2 requires the UI.X add-on. Frameworks can also hit rebuild limits (open_files_limit — community reports on UI.X). |
| Child-style workflow | Can I leave the vendor parent untouched? | If the vendor tells you to edit the parent, walk away. |
| Job fit | Magazine, gaming, “just a better Default,” wallpaper, docs. | The historical Empire / Dark Empire / Vertiforo piece is about jobs. Same test, living products. |
A June 2025 community thread still asks Bolt (PixelExit) versus UI.X 2. That is a real buyer question. It is not a winner. Bolt sits in the PixelExit / foundation family with XenBase. UI.X 2 sits in the ThemeHouse framework family. Pick the vendor whose update habit and add-on tax you can live with.
Snapshot: frameworks and foundations
UI.X 2 (ThemeHouse)
- Pitch (vendor): “Ultimate Theme Framework.”
- Updated: 25 June 2026 (RM snapshot).
- Requires: the UI.X add-on. Light + Dark.
- Job: a long-term parent you will child, with framework features the Default style does not pretend to have.
This is the “we are a studio and we will still be here next year” purchase. The add-on tax is real. Community practice: a UI.X rebuild can hit too many open files — raise open_files_limit (an example people use is 50000) or rebuild via CLI. That is an operations note, not a reason to avoid the product, and not a reason to buy it if you wanted a $30 skin.
If you do not want an extra add-on in the stack, UI.X 2 is the wrong column. Read XenBase or a DohTheme one-shot, or stay on stock variations.
XenBase (PixelExit)
- Pitch: a foundation style.
- Updated: 28 January 2026.
- Job: a parent you build on, in the same mental family as the Bolt question.
January 2026 is still inside the 2.3 era for this snapshot, but it is older than UI.X 2’s June 2026 bump and older than the DohTheme 29 June 2026 cluster. If your test is “did they ship after the last XF security/point release I care about,” check the RM on the day you buy. This article will not invent a compatibility matrix the listing did not print.
XenBase is the framework-shaped answer when you want PixelExit’s workflow rather than ThemeHouse’s. It is not automatically lighter. Foundations still want a child.
Snapshot: ThemeHouse Legend
- Job: gaming.
- Price listed: $59.50.
- Updated: 29 July 2026.
Legend is a one-job product with a fresh date and a public price. Compare it to stock-plus-variations only if your community is actually a gaming room that needs that chrome. Compare it to the old Vertiforo job (vertical / gaming) if you arrived from a historical Styles Factory URL: Legend is a thing you can still license. Vertiforo is not.
$59.50 is the RM figure in the snapshot. ThemeHouse can change it. The number is here so you have a scale next to the DohTheme $30–$55 cluster, not so you can quote it as a contract.
Snapshot: DohTheme line at 2.3.11
The DohTheme catalog on this snapshot is labelled 2.3.11. Many of the rows updated 29 June 2026. Prices are the RM list prices in that snapshot.
| Style | Price | Notes |
|---|---|---|
| Minimal | $40 | |
| Cozy | $30 | |
| Solfege | $55 | |
| Fury | $30 | |
| Alt | $30 | |
| Retro | $55 | |
| Pad | $30 | |
| Gamez | $30 | Gaming-shaped name; still a one-shot |
| Lotus | $30 | |
| Flow | $30 | |
| MaterialXen | $30 | |
| FluentXen | $30 | |
| Tiles | $30 | |
| Vigas | $30 | |
| xenAwsome | $30 | Spelling as listed |
| Mere | $30 |
These are one-shots: you buy a look. That is a legitimate job. The risk is the opposite of a framework: when XenForo 2.3.x moves a template, you need this vendor to ship 2.3.12, not a framework team plus your own child. The June 2026 cluster is a good sign at the time of the snapshot. It is not a promise.
Pick DohTheme when the demo is already 80% of what you would have painted in extra.less, and you are willing to live in a child of that parent. Do not pick four DohThemes and let members switch between them. That is four outdated-template queues. One parent, one child, stock variations if the product supports them.
Snapshot: newer 2026 names, and names you should treat as stale
Velqeris (Eggy) — dark purple / cyan, listed among the newer 2026 arrivals. Job: a finished dark aesthetic. Check the RM date on the day you buy; this snapshot names it as new, not as a framework.
Marmaris (erkanshi) — also in the newer-2026 set. Same buyer test: date, 2.3 variations, child-style workflow.
Uniform (Ehren) — wallpaper-focused. Updated 2 September 2024 on the snapshot. That date may lag 2.3. Say that out loud before you pay. A wallpaper style that has not moved since 2024 is a demo, not a parent, until the author ships.
Shades of Grey (Gator) — unmaintained since 2022. Do not install this on a 2.3 board and then file a ticket because dark mode “broke.” It is a historical artefact. If you like the look, copy the job (a restrained grey palette) into a child of Default with variations. Leave the 2022 templates alone.
Anything else on the RM with a last update in 2022–2023 gets the same sentence: it is a screenshot, not a product.
Add-on tax and rebuild cost (the part demos hide)
A style that requires a companion add-on has a second failure mode. The add-on must survive the next XF upgrade with the style. UI.X 2 is explicit about this. Other vendors sometimes ship “helpers” that are optional until you tick a demo feature.
Community-named add-ons in the same notebook as UI.X rebuilds — Redis Cache by Xon, Font Awesome Manager by Kirby, LiteSpeed Cache for XF2, XFOptimize — are stack tools, not themes. Do not buy them to fix a skin. XenPorta 2 is a portal add-on with known slider / lazy-load conflicts; it is not a style, and it is how people accidentally buy a second homepage. The portal article already told you to decide the first-screen job with stock tools first.
If a style demo leans on XenPorta, XFMG masonry, and three widget positions you do not have, the demo is a studio site. Subtract those before you fall in love.
UI.X rebuild hitting too many open files is the one concrete operations gotcha in the community notes. Raise the limit or rebuild from CLI. Do not lower memory_limit in a panic and call it a theme bug.
A buying sequence that does not waste a weekend
- Write the job in one sentence. “Support board that must look like our docs site.” “Gaming hub that should feel vertical and dark.” “Magazine front door.” If the sentence is “look modern,” you do not have a job. Stay on stock variations.
- Try the free baseline. Child of Default, variations on, palette honest, logo in. Live with it for a week on a staging board. The configuration-mistakes article already has the ticket titled “dark mode broke extra.less” — do not hard-code hex.
- Open the RM Styles [2.x] category. Filter with your eyes: last update in 2026, XF 2.3 mentioned, variations or Light+Dark in the description.
- Pick a column. Framework (UI.X 2 or XenBase / Bolt family) versus one-shot (DohTheme cluster, Legend, Velqeris, Marmaris).
- Read the add-on list on the resource page. If UI.X is required and you did not want an add-on, stop.
- Install on staging, as a parent. Create a child immediately. Never customize the imported parent.
- Test both variations, the style chooser if you still expose it, HTML email, the login page, the editor, and one third-party add-on template you actually use (the first thing a gaming style breaks is usually a shop or a stream widget).
- Check outdated templates after you install, before you customize. A parent that arrives with a pile of outdated templates is already behind.
- Only then put custom CSS in the child’s
extra.lessor in TMS.
If step 7 fails on email or the editor, that is a vendor problem, not an extra.less problem.
How to read an RM resource page
The Styles [2.x] listing is a catalog. The resource page is the contract. Read it in this order:
- Last update and the stated XF version. A 2.3.11 label with a 2026 date is the cluster you want. A 2.2 screenshot with a 2023 date is a screenshot.
- Add-ons required. UI.X 2 will say UI.X. Some one-shots quietly want a slider, a portal, or a font pack. If the required add-on is not something you already run, add it to the tax column or walk away.
- Variations / Light + Dark. If the author still describes two separately installed styles as the way to get dark mode, you are buying a 2.2 workflow. Prefer products that sit on 2.3 variations.
- Support thread recency. A resource with a June 2026 build and a silent support forum is still better than a 2022 build with a busy argument. You are buying merges, not vibes.
- Demo honesty. Subtract XFMG, XenPorta, and custom widgets. What remains is the style. If nothing remains, the style is a studio site.
Then install on staging anyway. Resource pages lie the way all storefronts lie: they show the day the author felt proud. Your editor and your oldest add-on are the real demo.
A staging test matrix (do not skip the boring rows)
Install is not QA. Walk this table on the child, on both variations, as a guest and as a Registered member.
| Surface | What “pass” looks like | Typical fail |
|---|---|---|
| Public index / your index route | Readable, no overlapping header, last-post column still a column | Framework header covering the sticky notice |
| Thread view + editor | Toolbar visible, quote works, attach works | Dark variation with white-on-white editor |
| Login / register / lost password | Forms contrast, no clipped buttons | Custom wrapper that only wrapped forum_list |
| HTML email | Default email style still readable | Public dark variation leaking into mail |
| Mobile width | Nav opens, guest CTA still tappable | Off-canvas that assumes a demo breakpoint |
| One real add-on template | The add-on you actually run (gallery, resources, a stream widget, a listing) | Demo never included it; you discover it in production |
| Appearance → Outdated templates | Empty, or only templates you chose to customize | Parent arrives already yellow |
Fail on email or the editor is a vendor problem. Fail on one add-on is either a child extra.less override or a style that was never tested against that add-on. Do not “fix” it by editing the vendor parent.
Re-parenting later is allowed: you can take a child of Default and point it at UI.X 2 or XenBase if you accept a merge. You cannot take a pile of parent edits and move them cleanly. That is why the child exists on day one.
What this snapshot is not
- Not a best-of list. UI.X 2 being updated in June 2026 does not make it better than XenBase for you. Legend at $59.50 is not “premium quality”; it is a listed price for a gaming job.
- Not historical Styles Factory products. Empire, Dark Empire, Vertiforo, FiveStyle are discontinued jobs. They are not on this RM table. Do not hunt a 2026 Empire version number. There isn’t one in this series.
- Not a light-versus-dark design guide. Variations are the mechanism. Whether your community should default dark is a later design article.
- Not an upgrade-safe customization manual. Child, TMS, merge — mentioned here so you do not skip them. The clicks live in the later customization article.
- Not official add-on pricing. XFMG, XFRM, and XFES are XenForo products. This article will not invent 2026 dollar prices for them.
Cloud versus self-hosted does not change the style model. It changes whether you can install a vendor add-on. Confirm that before you buy UI.X 2.
A short decision table
| If your job is… | First try | Then consider | Avoid |
|---|---|---|---|
| Clean support / docs board | Stock 2.3 variations + child | XenBase or a quiet DohTheme (Minimal, Mere, Flow) | Gaming skins, unmaintained greys |
| Magazine / portal chrome | Stock index route + widgets + article forum | A framework if the chrome is the product | Installing XenPorta because the demo used it |
| Gaming community | Stock dark variation | Legend ($59.50, Jul 2026), Gamez ($30), Vertiforo’s job on a living product | 2022 templates, four member-selectable skins |
| Long-term studio parent | — | UI.X 2 (Jun 2026, add-on tax) or XenBase (Jan 2026) | Editing the vendor parent |
| Wallpaper / art-heavy | Stock + a header image you own | Uniform only if the 2024 date is acceptable to you | Assuming it is 2.3-ready |
| “I liked this 2022 demo” | Recreate the palette on Default | — | Shades of Grey as a parent |
Checklist before you pay
- I wrote the job in one sentence.
- I lived on a child of Default with variations long enough to know what is still missing.
- The RM update date is 2026 (or I explicitly accepted an older date, as with Uniform).
- I know the add-on tax (UI.X 2 = UI.X add-on).
- I will install as a parent and customize only a child.
- I will not run two paid parents as a fake light/dark switch.
- I will put CSS in the child’s
extra.less(community) or TMS, not in Master. - I will not install Shades of Grey, or any 2022 parent, on 2.3 and call it a theme strategy.
Takeaways
- The best theme is the one whose job, cadence, and add-on tax you can defend. This page will not pick it for you.
- Stock 2.3 variations are the free baseline. A lot of “we need a theme” tickets die there.
- UI.X 2 (25 Jun 2026, framework + add-on, Light/Dark) and XenBase (28 Jan 2026, foundation) are the long-term parents in this snapshot. Bolt vs UI.X 2 is still a real question.
- Legend is $59.50 and current (29 Jul 2026) for a gaming job. DohTheme 2.3.11 is a $30–$55 one-shot cluster, many updated 29 Jun 2026.
- Velqeris and Marmaris are newer 2026 names. Uniform may lag 2.3 (Sep 2024). Shades of Grey has been unmaintained since 2022.
- Child the vendor parent. Do not edit Master.
extra.lessis community practice. Font file upload is not in the official docs.
When you outgrow this snapshot, the RM category is the source. Come back for the job test, not for a medal.