Organisation card: HTML with image cards; no compute section

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+images/logobrief.jpg filter=lfs diff=lfs merge=lfs -text

README.md

@@ -6,71 +6,85 @@ sdk: static
 pinned: false
 ---
 
-# LogoLabs
-
-LogoLabs designs custom logos and brand identities for startups, SMEs and founders, working
-from data. Our process is research-backed: we study how brands are drawn, what their briefs
-ask for and which choices hold up, and a short conversational brief becomes a set of original
-concepts, refined with you. The result is a logo and full brand guidelines in days, not weeks.
-
-The research behind that work is open. Here you'll find the models, tools and datasets we
-build it on: generating marks, turning them into clean vector files, and understanding what
-makes a design brief. The tools run on your own machine or in your browser, and nothing you
-draw or type is sent anywhere.
-
-**Want a finished identity rather than a tool?** We design custom logos with 20-page brand
-guidelines and full commercial rights, delivered in 24–48 hours, from €20. Start at
-[logolabs.org](https://logolabs.org) or [contact us](https://logolabs.org/contact).
-
-## Inkvec: raster to SVG, exactly
-
-[**Inkvec**](https://github.com/logolabs/inkvec) is an open-source (Apache-2.0) vectoriser
-that turns a PNG, JPEG or WebP logo into a clean SVG. It fits every shape by minimum
-description length: edges placed to a fraction of a pixel, the fewest curves that match
-the image, gradients kept as gradients, and transparency traced as it is. In our published
-comparison it has the lowest colour error on every case against VTracer and Trazor, and a
-Fast mode trades a little of that precision for speed.
-
-- **Try it in the browser:** [Inkvec Studio Lite](https://huggingface.co/spaces/Logolabs/inkvec),
-  the full editor as WebAssembly, with before-and-after comparisons on real logos.
-- **Desktop app and command line** for Windows, macOS and Linux:
-  [releases](https://github.com/logolabs/inkvec/releases).
-- **Bindings** for Python, JavaScript, Go, Swift, Java, .NET, PHP and C, all generated from
-  one options schema.
-
-Two small models ship with it, and are released on their own too:
-
-| Model | What it does | Size | Licence |
-|---|---|---|---|
-| [inkvec-denoiser-001](https://huggingface.co/Logolabs/inkvec-denoiser-001) | Removes JPEG, WebP and AI-decoder damage from flat artwork before tracing | 19.7M | Apache-2.0 |
-| [inkvec-sr-001](https://huggingface.co/Logolabs/inkvec-sr-001) | 4x super-resolution for logos and icons (MambaIRv2-Small, fine-tuned) | 9.77M | Apache-2.0 |
-
-## Agate: a compact text-to-image model
-
-[**Agate Preview 001**](https://huggingface.co/Logolabs/agate-preview-001) is a
-260M-parameter text-to-image model trained from scratch in 145 GPU-hours. It scores 0.550
-on GenEval with the official scorer, level with SDXL's published 0.55, and runs in under two
-seconds on a consumer GPU (MIT).
-
-- **Try it in the browser:** [Agate WebGPU](https://huggingface.co/spaces/Logolabs/agate-webgpu),
-  where the whole model runs on your own GPU; prompts never leave your machine.
-- **ComfyUI:** the [agate-comfyui](https://github.com/logolabs/agate-comfyui) node pack.
-
-## LogoBrief-10K: a dataset of logos and their briefs
-
-[**LogoBrief-10K**](https://huggingface.co/datasets/Logolabs/LogoBrief-10K) is 10,000 brand
-logos, each with its original SVG and a clean raster render, plus design annotations:
-open-vocabulary style tags, a one-sentence motif description, a full generated design brief,
-and text-region bounding boxes. Domains are sampled by web-popularity rank rather than picked
-for recognisable brands, and every one was checked for AI-training opt-out signals before
-publication. Access is gated; the card's preview images are in
-[LogoBrief-10K-assets](https://huggingface.co/datasets/Logolabs/LogoBrief-10K-assets).
-
-## Compute
-
-Our models were trained on **Arrhenius** at NAISS, through **EuroHPC JU**.
-
-## Links
-
-[logolabs.org](https://logolabs.org) · [Contact](https://logolabs.org/contact) ·
-[GitHub](https://github.com/logolabs) · [Inkvec documentation](https://logolabs.github.io/inkvec)
+<style>
+  .ll-lead { font-size: 1.05rem; line-height: 1.65; }
+  .ll-copper { color: #c9754a; }
+  .ll-btn { display: inline-block; padding: .45rem .95rem; border-radius: .6rem; font-weight: 600; font-size: .9rem; text-decoration: none !important; }
+  .ll-btn-solid { background: #c9754a; color: #fff !important; }
+  .ll-btn-solid:hover { background: #b5643c; }
+  .ll-btn-line { border: 1px solid #c9754a; color: #c9754a !important; }
+  .ll-card { display: block; border-radius: .9rem; overflow: hidden; text-decoration: none !important; transition: transform .15s ease; }
+  .ll-card:hover { transform: translateY(-2px); }
+  .ll-card img { width: 100%; height: 11rem; object-fit: cover; display: block; margin: 0; }
+  .ll-kicker { font-size: .72rem; letter-spacing: .08em; text-transform: uppercase; color: #c9754a; font-weight: 600; }
+  .ll-links a { color: #c9754a; }
+</style>
+
+<p class="ll-lead mb-3">
+  <b>LogoLabs designs custom logos and brand identities from data.</b>
+  We study how brands are drawn, what their briefs ask for and which choices hold up, and turn a short
+  conversational brief into original concepts, refined with you: a logo and full brand guidelines in days, not weeks.
+</p>
+
+<p class="mb-5 text-gray-500 dark:text-gray-400">
+  The research behind that work is open. These are the models, tools and data we build it on,
+  and they run on your own machine or in your browser.
+</p>
+
+<div class="flex flex-wrap gap-3 mb-8">
+  <a class="ll-btn ll-btn-solid" href="https://logolabs.org" target="_blank">Get a logo designed, from €20</a>
+  <a class="ll-btn ll-btn-line" href="https://logolabs.org/contact" target="_blank">Contact us</a>
+</div>
+
+<div class="grid md:grid-cols-3 gap-5 mb-8">
+
+  <a class="ll-card border border-gray-200 dark:border-gray-800" href="https://huggingface.co/spaces/Logolabs/inkvec">
+    <img src="images/inkvec.jpg" alt="Inkvec before and after, with the traced anchors" />
+    <div class="p-4">
+      <div class="ll-kicker mb-1">Vectoriser · Apache-2.0</div>
+      <div class="font-semibold text-lg mb-1">Inkvec</div>
+      <div class="text-sm text-gray-500 dark:text-gray-400">
+        Turns a PNG, JPEG or WebP logo into a clean SVG: edges to a fraction of a pixel, the fewest curves that match,
+        gradients and transparency kept. Try Studio Lite in the browser.
+      </div>
+    </div>
+  </a>
+
+  <a class="ll-card border border-gray-200 dark:border-gray-800" href="https://huggingface.co/Logolabs/agate-preview-001">
+    <img src="images/agate.jpg" alt="Agate model release" />
+    <div class="p-4">
+      <div class="ll-kicker mb-1">Text to image · MIT</div>
+      <div class="font-semibold text-lg mb-1">Agate Preview 001</div>
+      <div class="text-sm text-gray-500 dark:text-gray-400">
+        A 260M-parameter model trained from scratch in 145 GPU-hours. GenEval 0.550, level with SDXL,
+        and it runs in your browser on WebGPU.
+      </div>
+    </div>
+  </a>
+
+  <a class="ll-card border border-gray-200 dark:border-gray-800" href="https://huggingface.co/datasets/Logolabs/LogoBrief-10K">
+    <img src="images/logobrief.jpg" alt="Logos from LogoBrief-10K" />
+    <div class="p-4">
+      <div class="ll-kicker mb-1">Dataset · gated</div>
+      <div class="font-semibold text-lg mb-1">LogoBrief-10K</div>
+      <div class="text-sm text-gray-500 dark:text-gray-400">
+        10,000 brand logos with their original SVGs, style tags, motif descriptions and full design briefs,
+        sampled by web popularity and checked for AI-training opt-outs.
+      </div>
+    </div>
+  </a>
+
+</div>
+
+<p class="text-sm text-gray-500 dark:text-gray-400 mb-2">
+  Also: <a class="ll-copper" href="https://huggingface.co/spaces/Logolabs/agate-webgpu">Agate WebGPU</a> ·
+  <a class="ll-copper" href="https://github.com/logolabs/agate-comfyui">Agate for ComfyUI</a> ·
+  <a class="ll-copper" href="https://huggingface.co/Logolabs/inkvec-denoiser-001">inkvec-denoiser-001</a> (JPEG and AI-damage cleanup) ·
+  <a class="ll-copper" href="https://huggingface.co/Logolabs/inkvec-sr-001">inkvec-sr-001</a> (4× upscaling for logos) ·
+  <a class="ll-copper" href="https://github.com/logolabs/inkvec/releases">Inkvec desktop app</a>
+</p>
+
+<p class="ll-links text-sm text-gray-500 dark:text-gray-400">
+  <a href="https://logolabs.org">logolabs.org</a> · <a href="https://logolabs.org/contact">Contact</a> ·
+  <a href="https://github.com/logolabs">GitHub</a> · <a href="https://logolabs.github.io/inkvec">Inkvec docs</a>
+</p>

index.html

@@ -1,92 +1,102 @@
 <!doctype html>
-<html lang="en">
+<html lang="en" class="dark">
 <head>
 <meta charset="utf-8">
 <meta name="viewport" content="width=device-width, initial-scale=1">
 <title>LogoLabs</title>
-<meta name="description" content="LogoLabs designs custom logos and brand identities from data, and publishes the research behind it: Inkvec, Agate and LogoBrief-10K.">
-<link rel="preconnect" href="https://fonts.googleapis.com">
-<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
-<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600&family=Playfair+Display:wght@600;700&display=swap" rel="stylesheet">
-<link rel="stylesheet" href="style.css">
+<meta name="description" content="LogoLabs designs custom logos and brand identities from data, and publishes the research behind it.">
+<script src="https://cdn.tailwindcss.com"></script>
+<script>tailwind.config = { darkMode: 'class' }</script>
+<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;600&family=Playfair+Display:wght@700&display=swap" rel="stylesheet">
+<style>body { font-family: Inter, system-ui, sans-serif; } h1 { font-family: 'Playfair Display', Georgia, serif; }</style>
 </head>
-<body>
-<main>
-<h1>LogoLabs</h1>
-<p>LogoLabs designs custom logos and brand identities for startups, SMEs and founders, working
-from data. Our process is research-backed: we study how brands are drawn, what their briefs
-ask for and which choices hold up, and a short conversational brief becomes a set of original
-concepts, refined with you. The result is a logo and full brand guidelines in days, not weeks.</p>
-<p>The research behind that work is open. Here you'll find the models, tools and datasets we
-build it on: generating marks, turning them into clean vector files, and understanding what
-makes a design brief. The tools run on your own machine or in your browser, and nothing you
-draw or type is sent anywhere.</p>
-<p><strong>Want a finished identity rather than a tool?</strong> We design custom logos with 20-page brand
-guidelines and full commercial rights, delivered in 24–48 hours, from €20. Start at
-<a href="https://logolabs.org" target="_blank" rel="noopener">logolabs.org</a> or <a href="https://logolabs.org/contact" target="_blank" rel="noopener">contact us</a>.</p>
-<h2>Inkvec: raster to SVG, exactly</h2>
-<p><a href="https://github.com/logolabs/inkvec" target="_blank" rel="noopener"><strong>Inkvec</strong></a> is an open-source (Apache-2.0) vectoriser
-that turns a PNG, JPEG or WebP logo into a clean SVG. It fits every shape by minimum
-description length: edges placed to a fraction of a pixel, the fewest curves that match
-the image, gradients kept as gradients, and transparency traced as it is. In our published
-comparison it has the lowest colour error on every case against VTracer and Trazor, and a
-Fast mode trades a little of that precision for speed.</p>
-<ul>
-<li><strong>Try it in the browser:</strong> <a href="https://huggingface.co/spaces/Logolabs/inkvec" target="_blank" rel="noopener">Inkvec Studio Lite</a>,
-  the full editor as WebAssembly, with before-and-after comparisons on real logos.</li>
-<li><strong>Desktop app and command line</strong> for Windows, macOS and Linux:
-  <a href="https://github.com/logolabs/inkvec/releases" target="_blank" rel="noopener">releases</a>.</li>
-<li><strong>Bindings</strong> for Python, JavaScript, Go, Swift, Java, .NET, PHP and C, all generated from
-  one options schema.</li>
-</ul>
-<p>Two small models ship with it, and are released on their own too:</p>
-<table>
-<thead>
-<tr>
-<th>Model</th>
-<th>What it does</th>
-<th>Size</th>
-<th>Licence</th>
-</tr>
-</thead>
-<tbody>
-<tr>
-<td><a href="https://huggingface.co/Logolabs/inkvec-denoiser-001" target="_blank" rel="noopener">inkvec-denoiser-001</a></td>
-<td>Removes JPEG, WebP and AI-decoder damage from flat artwork before tracing</td>
-<td>19.7M</td>
-<td>Apache-2.0</td>
-</tr>
-<tr>
-<td><a href="https://huggingface.co/Logolabs/inkvec-sr-001" target="_blank" rel="noopener">inkvec-sr-001</a></td>
-<td>4x super-resolution for logos and icons (MambaIRv2-Small, fine-tuned)</td>
-<td>9.77M</td>
-<td>Apache-2.0</td>
-</tr>
-</tbody>
-</table>
-<h2>Agate: a compact text-to-image model</h2>
-<p><a href="https://huggingface.co/Logolabs/agate-preview-001" target="_blank" rel="noopener"><strong>Agate Preview 001</strong></a> is a
-260M-parameter text-to-image model trained from scratch in 145 GPU-hours. It scores 0.550
-on GenEval with the official scorer, level with SDXL's published 0.55, and runs in under two
-seconds on a consumer GPU (MIT).</p>
-<ul>
-<li><strong>Try it in the browser:</strong> <a href="https://huggingface.co/spaces/Logolabs/agate-webgpu" target="_blank" rel="noopener">Agate WebGPU</a>,
-  where the whole model runs on your own GPU; prompts never leave your machine.</li>
-<li><strong>ComfyUI:</strong> the <a href="https://github.com/logolabs/agate-comfyui" target="_blank" rel="noopener">agate-comfyui</a> node pack.</li>
-</ul>
-<h2>LogoBrief-10K: a dataset of logos and their briefs</h2>
-<p><a href="https://huggingface.co/datasets/Logolabs/LogoBrief-10K" target="_blank" rel="noopener"><strong>LogoBrief-10K</strong></a> is 10,000 brand
-logos, each with its original SVG and a clean raster render, plus design annotations:
-open-vocabulary style tags, a one-sentence motif description, a full generated design brief,
-and text-region bounding boxes. Domains are sampled by web-popularity rank rather than picked
-for recognisable brands, and every one was checked for AI-training opt-out signals before
-publication. Access is gated; the card's preview images are in
-<a href="https://huggingface.co/datasets/Logolabs/LogoBrief-10K-assets" target="_blank" rel="noopener">LogoBrief-10K-assets</a>.</p>
-<h2>Compute</h2>
-<p>Our models were trained on <strong>Arrhenius</strong> at NAISS, through <strong>EuroHPC JU</strong>.</p>
-<h2>Links</h2>
-<p><a href="https://logolabs.org" target="_blank" rel="noopener">logolabs.org</a> · <a href="https://logolabs.org/contact" target="_blank" rel="noopener">Contact</a> ·
-<a href="https://github.com/logolabs" target="_blank" rel="noopener">GitHub</a> · <a href="https://logolabs.github.io/inkvec" target="_blank" rel="noopener">Inkvec documentation</a></p>
+<body class="bg-[#141414] text-gray-100">
+<main class="max-w-5xl mx-auto px-5 py-12">
+<h1 class="text-4xl mb-6">LogoLabs</h1>
+
+<style>
+  .ll-lead { font-size: 1.05rem; line-height: 1.65; }
+  .ll-copper { color: #c9754a; }
+  .ll-btn { display: inline-block; padding: .45rem .95rem; border-radius: .6rem; font-weight: 600; font-size: .9rem; text-decoration: none !important; }
+  .ll-btn-solid { background: #c9754a; color: #fff !important; }
+  .ll-btn-solid:hover { background: #b5643c; }
+  .ll-btn-line { border: 1px solid #c9754a; color: #c9754a !important; }
+  .ll-card { display: block; border-radius: .9rem; overflow: hidden; text-decoration: none !important; transition: transform .15s ease; }
+  .ll-card:hover { transform: translateY(-2px); }
+  .ll-card img { width: 100%; height: 11rem; object-fit: cover; display: block; margin: 0; }
+  .ll-kicker { font-size: .72rem; letter-spacing: .08em; text-transform: uppercase; color: #c9754a; font-weight: 600; }
+  .ll-links a { color: #c9754a; }
+</style>
+
+<p class="ll-lead mb-3">
+  <b>LogoLabs designs custom logos and brand identities from data.</b>
+  We study how brands are drawn, what their briefs ask for and which choices hold up, and turn a short
+  conversational brief into original concepts, refined with you: a logo and full brand guidelines in days, not weeks.
+</p>
+
+<p class="mb-5 text-gray-500 dark:text-gray-400">
+  The research behind that work is open. These are the models, tools and data we build it on,
+  and they run on your own machine or in your browser.
+</p>
+
+<div class="flex flex-wrap gap-3 mb-8">
+  <a class="ll-btn ll-btn-solid" href="https://logolabs.org" target="_blank">Get a logo designed, from €20</a>
+  <a class="ll-btn ll-btn-line" href="https://logolabs.org/contact" target="_blank">Contact us</a>
+</div>
+
+<div class="grid md:grid-cols-3 gap-5 mb-8">
+
+  <a class="ll-card border border-gray-200 dark:border-gray-800" href="https://huggingface.co/spaces/Logolabs/inkvec">
+    <img src="images/inkvec.jpg" alt="Inkvec before and after, with the traced anchors" />
+    <div class="p-4">
+      <div class="ll-kicker mb-1">Vectoriser · Apache-2.0</div>
+      <div class="font-semibold text-lg mb-1">Inkvec</div>
+      <div class="text-sm text-gray-500 dark:text-gray-400">
+        Turns a PNG, JPEG or WebP logo into a clean SVG: edges to a fraction of a pixel, the fewest curves that match,
+        gradients and transparency kept. Try Studio Lite in the browser.
+      </div>
+    </div>
+  </a>
+
+  <a class="ll-card border border-gray-200 dark:border-gray-800" href="https://huggingface.co/Logolabs/agate-preview-001">
+    <img src="images/agate.jpg" alt="Agate model release" />
+    <div class="p-4">
+      <div class="ll-kicker mb-1">Text to image · MIT</div>
+      <div class="font-semibold text-lg mb-1">Agate Preview 001</div>
+      <div class="text-sm text-gray-500 dark:text-gray-400">
+        A 260M-parameter model trained from scratch in 145 GPU-hours. GenEval 0.550, level with SDXL,
+        and it runs in your browser on WebGPU.
+      </div>
+    </div>
+  </a>
+
+  <a class="ll-card border border-gray-200 dark:border-gray-800" href="https://huggingface.co/datasets/Logolabs/LogoBrief-10K">
+    <img src="images/logobrief.jpg" alt="Logos from LogoBrief-10K" />
+    <div class="p-4">
+      <div class="ll-kicker mb-1">Dataset · gated</div>
+      <div class="font-semibold text-lg mb-1">LogoBrief-10K</div>
+      <div class="text-sm text-gray-500 dark:text-gray-400">
+        10,000 brand logos with their original SVGs, style tags, motif descriptions and full design briefs,
+        sampled by web popularity and checked for AI-training opt-outs.
+      </div>
+    </div>
+  </a>
+
+</div>
+
+<p class="text-sm text-gray-500 dark:text-gray-400 mb-2">
+  Also: <a class="ll-copper" href="https://huggingface.co/spaces/Logolabs/agate-webgpu">Agate WebGPU</a> ·
+  <a class="ll-copper" href="https://github.com/logolabs/agate-comfyui">Agate for ComfyUI</a> ·
+  <a class="ll-copper" href="https://huggingface.co/Logolabs/inkvec-denoiser-001">inkvec-denoiser-001</a> (JPEG and AI-damage cleanup) ·
+  <a class="ll-copper" href="https://huggingface.co/Logolabs/inkvec-sr-001">inkvec-sr-001</a> (4× upscaling for logos) ·
+  <a class="ll-copper" href="https://github.com/logolabs/inkvec/releases">Inkvec desktop app</a>
+</p>
+
+<p class="ll-links text-sm text-gray-500 dark:text-gray-400">
+  <a href="https://logolabs.org">logolabs.org</a> · <a href="https://logolabs.org/contact">Contact</a> ·
+  <a href="https://github.com/logolabs">GitHub</a> · <a href="https://logolabs.github.io/inkvec">Inkvec docs</a>
+</p>
+
 </main>
 </body>
 </html>

style.css

@@ -1,20 +0,0 @@
-/* LogoLabs: charcoal and copper, Playfair Display for headings, Inter for text. */
-:root { --bg: #141414; --panel: #1c1c1c; --rule: #2e2e2e; --text: #e8e4de; --dim: #a8a29a; --copper: #c9754a; }
-* { box-sizing: border-box; }
-body { margin: 0; background: var(--bg); color: var(--text); font: 16px/1.65 Inter, system-ui, sans-serif; }
-main { max-width: 820px; margin: 0 auto; padding: 56px 20px 80px; }
-h1, h2 { font-family: "Playfair Display", Georgia, serif; line-height: 1.2; }
-h1 { font-size: 44px; margin: 0 0 20px; }
-h1::after { content: ""; display: block; width: 56px; height: 3px; background: var(--copper); margin-top: 14px; }
-h2 { font-size: 26px; margin: 44px 0 12px; padding-top: 20px; border-top: 1px solid var(--rule); }
-p, li { color: var(--text); }
-strong { color: #fff; font-weight: 600; }
-a { color: var(--copper); text-decoration: none; border-bottom: 1px solid rgba(201,117,74,.35); }
-a:hover { border-bottom-color: var(--copper); }
-ul { padding-left: 20px; }
-li { margin: 6px 0; }
-table { width: 100%; border-collapse: collapse; margin: 16px 0; font-size: 14.5px; background: var(--panel); border-radius: 8px; overflow: hidden; }
-th, td { text-align: left; padding: 10px 12px; border-bottom: 1px solid var(--rule); vertical-align: top; }
-th { color: var(--dim); font-weight: 500; font-size: 13px; text-transform: uppercase; letter-spacing: .04em; }
-tr:last-child td { border-bottom: 0; }
-@media (max-width: 600px) { h1 { font-size: 34px; } table { display: block; overflow-x: auto; } }