Edit photos by describing the change, on your own GPU. A free, open-source alternative to Nano Banana and ChatGPT image editing: photos stay on your computer, no daily limit. One installer for Windows, Mac and Linux sets up the best open model (Qwen-Image 2.1). MCP server for AI agents.
TypeScript ai-photo-editorcomfyui-alternativegenerative-fillimage-editingimage-generationinpaintinglinuxlocal-aimacosmcpmcp-servernano-banana-alternativeofflinephoto-editorprivacyqwen-imageqwen-image-2-1stable-diffusion-cppvulkanwindows
Website SEO: search-focused titles, structured data, sitemap, two guides
Gabriele Lozuponecb2fc95f184c
diff --git a/README.md b/README.mdindex dbb768d..081643f 100644--- a/README.md+++ b/README.md@@ -29,6 +29,7 @@</p><p align="center">+ <a href="https://gabrielelozupone.github.io/kokon/">Website</a> ·<a href="#install">Install</a> ·<a href="#what-you-can-do">Features</a> ·<a href="#examples">Examples</a> ·diff --git a/site/assets/img/app-screenshot.webp b/site/assets/img/app-screenshot.webpnew file mode 100644index 0000000..67290ceBinary files /dev/null and b/site/assets/img/app-screenshot.webp differdiff --git a/site/assets/style.css b/site/assets/style.cssnew file mode 100644index 0000000..5efe3bc--- /dev/null+++ b/site/assets/style.css@@ -0,0 +1,245 @@+:root {+ --bg: #070707;+ --bg-2: #0d0d0d;+ --card: rgba(255, 255, 255, 0.035);+ --card-hi: rgba(255, 255, 255, 0.06);+ --line: rgba(255, 255, 255, 0.08);+ --line-hi: rgba(255, 255, 255, 0.14);+ --text: #f2f1ed;+ --text-2: rgba(242, 241, 237, 0.66);+ --text-3: rgba(242, 241, 237, 0.42);+ --paper: #f7f6f2;+ --ink: #0d0d0d;+ --radius: 22px;+ --max: 1200px;+ --font: "Inter", system-ui, -apple-system, "Segoe UI", sans-serif;+ --mono: "JetBrains Mono", "Cascadia Code", Consolas, monospace;+ color-scheme: dark;+}+* { box-sizing: border-box; }+html { scroll-behavior: smooth; -webkit-text-size-adjust: 100%; }+body { margin: 0; background: var(--bg); color: var(--text); font-family: var(--font); font-size: 17px; line-height: 1.55; -webkit-font-smoothing: antialiased; overflow-x: hidden; }+img, video { display: block; max-width: 100%; }+a { color: inherit; text-decoration: none; }+button { font: inherit; color: inherit; background: none; border: 0; cursor: pointer; padding: 0; }+.wrap { width: min(var(--max), calc(100% - 32px)); margin: 0 auto; }+@media (min-width: 720px) { .wrap { width: min(var(--max), calc(100% - 64px)); } }++/* Grain + light */+body::before { content: ""; position: fixed; inset: 0; pointer-events: none; z-index: 100; opacity: 0.035; background-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='160' height='160'%3E%3Cfilter id='n'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.9' numOctaves='2' stitchTiles='stitch'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' filter='url(%23n)'/%3E%3C/svg%3E"); }++/* Nav */+.nav { position: fixed; top: 0; left: 0; right: 0; z-index: 50; transition: background 0.3s, border-color 0.3s, backdrop-filter 0.3s; border-bottom: 1px solid transparent; }+.nav.scrolled { background: rgba(7, 7, 7, 0.72); backdrop-filter: saturate(1.4) blur(18px); -webkit-backdrop-filter: saturate(1.4) blur(18px); border-color: var(--line); }+.nav .wrap { height: 64px; display: flex; align-items: center; gap: 28px; }+.logo { display: flex; align-items: center; gap: 10px; font-weight: 640; font-size: 19px; letter-spacing: -0.03em; }+.logo svg { width: 22px; height: 22px; }+.nav-links { display: none; gap: 26px; font-size: 14.5px; color: var(--text-2); }+.nav-links a:hover { color: var(--text); }+.nav .spacer { flex: 1; }+.gh { display: none; align-items: center; gap: 8px; font-size: 14px; color: var(--text-2); }+.gh:hover { color: var(--text); }+@media (min-width: 900px) { .nav-links, .gh { display: flex; } }++.btn { display: inline-flex; align-items: center; justify-content: center; gap: 10px; height: 48px; padding: 0 22px; border-radius: 999px; font-weight: 560; font-size: 15.5px; letter-spacing: -0.01em; transition: transform 0.2s, background 0.2s, box-shadow 0.2s; white-space: nowrap; }+.btn svg { width: 18px; height: 18px; flex: none; }+.btn.primary { background: var(--paper); color: var(--ink); box-shadow: 0 0 0 1px rgba(255,255,255,0.2), 0 10px 40px rgba(247,246,242,0.12); }+.btn.primary:hover { transform: translateY(-1px); box-shadow: 0 0 0 1px rgba(255,255,255,0.3), 0 14px 50px rgba(247,246,242,0.2); }+.btn.ghost { background: var(--card-hi); box-shadow: inset 0 0 0 1px var(--line-hi); }+.btn.ghost:hover { background: rgba(255,255,255,0.1); }+.btn.small { height: 38px; padding: 0 16px; font-size: 14px; }+.btn .tag { font-size: 11.5px; font-weight: 600; padding: 2px 7px; border-radius: 6px; background: rgba(255,255,255,0.1); color: var(--text-2); letter-spacing: 0.02em; }++/* Hero */+.hero { position: relative; padding: 148px 0 40px; text-align: center; }+.hero::before { content: ""; position: absolute; left: 50%; top: -240px; width: 1400px; height: 900px; transform: translateX(-50%); background: radial-gradient(ellipse at center, rgba(247,246,242,0.09) 0%, rgba(247,246,242,0) 60%); pointer-events: none; }+.eyebrow { display: inline-flex; align-items: center; gap: 10px; padding: 6px 14px 6px 6px; border-radius: 999px; background: var(--card); box-shadow: inset 0 0 0 1px var(--line); font-size: 13.5px; color: var(--text-2); }+.eyebrow b { font-weight: 600; color: var(--ink); background: var(--paper); padding: 3px 9px; border-radius: 999px; font-size: 12px; }+h1 { font-size: clamp(44px, 8.4vw, 104px); line-height: 1; letter-spacing: -0.045em; font-weight: 680; margin: 26px auto 0; max-width: 11.5em; }+h1 .soft { color: var(--text-3); }+.lede { font-size: clamp(17px, 2vw, 21px); color: var(--text-2); max-width: 36em; margin: 26px auto 0; letter-spacing: -0.012em; }+.ctas { display: flex; flex-wrap: wrap; gap: 12px; justify-content: center; margin-top: 36px; }+.fine { margin-top: 18px; font-size: 13.5px; color: var(--text-3); }++.showreel { position: relative; margin: 72px auto 0; max-width: 1120px; }+.showreel .frame { position: relative; border-radius: 22px; overflow: hidden; background: #000 url("assets/kokon-poster.jpg") center / cover; box-shadow: 0 0 0 1px var(--line-hi), 0 60px 160px rgba(0,0,0,0.75), 0 0 120px rgba(247,246,242,0.06); aspect-ratio: 16 / 9; }+.showreel video { width: 100%; height: 100%; object-fit: cover; }+.showreel .glow { position: absolute; inset: 8% 4% -6%; background: radial-gradient(ellipse at center, rgba(255,214,140,0.16), transparent 65%); filter: blur(40px); z-index: -1; }+.sound { position: absolute; right: 16px; bottom: 16px; height: 36px; padding: 0 14px; border-radius: 999px; background: rgba(0,0,0,0.55); backdrop-filter: blur(10px); font-size: 13px; color: #fff; display: inline-flex; align-items: center; gap: 8px; box-shadow: inset 0 0 0 1px rgba(255,255,255,0.15); }++/* Runs on */+.runs { padding: 64px 0 24px; text-align: center; }+.runs p { font-size: 13px; letter-spacing: 0.16em; text-transform: uppercase; color: var(--text-3); margin: 0 0 22px; }+.runs ul { list-style: none; margin: 0; padding: 0; display: flex; flex-wrap: wrap; justify-content: center; gap: 14px 44px; }+.runs li { display: flex; align-items: center; gap: 10px; color: var(--text-2); font-weight: 560; font-size: 17px; letter-spacing: -0.01em; }+.runs svg { width: 24px; height: 24px; fill: currentColor; }++/* Sections */+section { position: relative; }+.sec { padding: 120px 0 0; }+.kicker { font-size: 13px; letter-spacing: 0.16em; text-transform: uppercase; color: var(--text-3); margin: 0 0 16px; }+h2 { font-size: clamp(36px, 5.6vw, 68px); line-height: 1.04; letter-spacing: -0.042em; font-weight: 670; margin: 0; max-width: 14em; }+.sec-lede { font-size: 19px; color: var(--text-2); max-width: 34em; margin: 20px 0 0; letter-spacing: -0.01em; }+.center { text-align: center; }+.center h2, .center .sec-lede { margin-left: auto; margin-right: auto; }++/* Compare */+.studio { margin-top: 48px; border-radius: 28px; background: linear-gradient(180deg, rgba(255,255,255,0.05), rgba(255,255,255,0.015)); box-shadow: inset 0 0 0 1px var(--line); padding: 14px; }+.tabs { display: flex; gap: 4px; overflow-x: auto; scrollbar-width: none; padding: 4px; background: rgba(0,0,0,0.35); border-radius: 16px; width: max-content; max-width: 100%; margin: 0 auto 14px; }+.tabs::-webkit-scrollbar { display: none; }+.tab { height: 38px; padding: 0 16px; border-radius: 12px; font-size: 14.5px; font-weight: 520; color: var(--text-2); white-space: nowrap; transition: background 0.2s, color 0.2s; }+.tab:hover { color: var(--text); }+.tab.on { background: var(--paper); color: var(--ink); }+.cmp { position: relative; border-radius: 16px; overflow: hidden; background: #000; user-select: none; touch-action: pan-y; cursor: ew-resize; aspect-ratio: 16 / 9; }+.cmp img { position: absolute; inset: 0; width: 100%; height: 100%; object-fit: contain; pointer-events: none; }+.cmp .before { clip-path: inset(0 calc(100% - var(--split, 50%)) 0 0); }+.cmp .handle { position: absolute; top: 0; bottom: 0; left: var(--split, 50%); width: 2px; margin-left: -1px; background: #fff; box-shadow: 0 0 0 1px rgba(0,0,0,0.25), 0 0 24px rgba(255,255,255,0.4); pointer-events: none; }+.cmp .handle span { position: absolute; top: 50%; left: 50%; width: 44px; height: 44px; margin: -22px 0 0 -22px; border-radius: 50%; background: #fff; box-shadow: 0 4px 18px rgba(0,0,0,0.45); display: grid; place-items: center; color: #111; }+.cmp .lbl { position: absolute; top: 14px; padding: 4px 10px; border-radius: 8px; font-size: 12.5px; color: #fff; background: rgba(0,0,0,0.55); backdrop-filter: blur(8px); pointer-events: none; }+.cmp .lbl.l { left: 14px; } .cmp .lbl.r { right: 14px; }+.cmp .bg-blur { position: absolute; inset: -40px; width: calc(100% + 80px); height: calc(100% + 80px); object-fit: cover; filter: blur(40px) brightness(0.35) saturate(1.2); transform: scale(1.1); }+.promptbar { margin-top: 14px; display: flex; align-items: center; gap: 14px; padding: 14px 14px 14px 20px; border-radius: 18px; background: #1a1a1a; box-shadow: inset 0 0 0 1px var(--line); }+.promptbar .mode { flex: none; font-size: 12.5px; font-weight: 560; padding: 4px 10px; border-radius: 8px; background: rgba(255,255,255,0.08); color: var(--text-2); }+.promptbar .txt { flex: 1; min-width: 0; font-size: 16px; color: var(--text); letter-spacing: -0.01em; }+.promptbar .send { flex: none; width: 38px; height: 38px; border-radius: 50%; background: var(--paper); color: var(--ink); display: grid; place-items: center; }+.honest { margin-top: 18px; text-align: center; font-size: 13.5px; color: var(--text-3); }++/* Bento */+.bento { margin-top: 56px; display: grid; gap: 16px; grid-template-columns: 1fr; }+@media (min-width: 900px) { .bento { grid-template-columns: repeat(6, 1fr); } .b-wide { grid-column: span 4; } .b-mid { grid-column: span 3; } .b-small { grid-column: span 2; } }+.card { position: relative; border-radius: var(--radius); background: linear-gradient(180deg, rgba(255,255,255,0.05), rgba(255,255,255,0.018)); box-shadow: inset 0 0 0 1px var(--line); padding: 30px; overflow: hidden; min-height: 280px; display: flex; flex-direction: column; }+.card h3 { margin: 0; font-size: 24px; letter-spacing: -0.035em; font-weight: 640; line-height: 1.15; }+.card p { margin: 10px 0 0; color: var(--text-2); font-size: 15.5px; max-width: 30em; }+.card .vis { margin-top: auto; padding-top: 26px; }+.icon { width: 40px; height: 40px; border-radius: 12px; background: rgba(255,255,255,0.07); display: grid; place-items: center; margin-bottom: 20px; box-shadow: inset 0 0 0 1px var(--line); }+.icon svg { width: 20px; height: 20px; stroke: var(--text); fill: none; stroke-width: 1.8; stroke-linecap: round; stroke-linejoin: round; }+.mask-demo { position: relative; border-radius: 14px; overflow: hidden; aspect-ratio: 16 / 9; }+.mask-demo img { width: 100%; height: 100%; object-fit: cover; }+.mask-demo svg { position: absolute; inset: 0; width: 100%; height: 100%; mix-blend-mode: difference; opacity: 0.75; }+.mask-demo .path { stroke-dasharray: 3000; stroke-dashoffset: 3000; transition: stroke-dashoffset 2.6s cubic-bezier(0.65, 0, 0.35, 1); }+.mask-demo.in .path { stroke-dashoffset: 0; }+.mask-demo .after { position: absolute; inset: 0; opacity: 0; transition: opacity 0.9s 2.8s; }+.mask-demo.in .after { opacity: 1; }+.chips { display: flex; flex-wrap: wrap; gap: 8px; }+.chip { padding: 8px 14px; border-radius: 999px; font-size: 13.5px; background: rgba(255,255,255,0.06); box-shadow: inset 0 0 0 1px var(--line); color: var(--text-2); }+.chip.on { background: var(--paper); color: var(--ink); box-shadow: none; font-weight: 560; }+.timeline { display: flex; align-items: center; gap: 0; }+.timeline .t { width: 64px; height: 64px; border-radius: 14px; overflow: hidden; flex: none; outline: 2px solid transparent; outline-offset: 3px; }+.timeline .t.sel { outline-color: var(--text); }+.timeline .t img { width: 100%; height: 100%; object-fit: cover; }+.timeline .j { width: 22px; height: 2px; background: var(--line-hi); flex: none; }+.timeline .lbls { display: flex; gap: 0; margin-top: 10px; font-size: 12px; color: var(--text-3); }+.big-num { font-size: 64px; font-weight: 680; letter-spacing: -0.05em; line-height: 1; }+.big-num small { font-size: 20px; font-weight: 520; color: var(--text-2); letter-spacing: -0.01em; margin-left: 6px; }++/* Agent */+.split { display: grid; gap: 48px; align-items: center; grid-template-columns: 1fr; }+@media (min-width: 980px) { .split { grid-template-columns: 1fr 1.12fr; gap: 72px; } }+.term { border-radius: 18px; background: #0c0c0c; box-shadow: 0 0 0 1px var(--line-hi), 0 40px 120px rgba(0,0,0,0.6); overflow: hidden; font-family: var(--mono); font-size: 13.5px; line-height: 1.7; }+.term-head { height: 40px; display: flex; align-items: center; gap: 8px; padding: 0 16px; background: #151515; font-family: var(--font); font-size: 12.5px; color: var(--text-3); }+.term-head i { width: 11px; height: 11px; border-radius: 50%; background: #333; }+.term-body { padding: 20px 22px 22px; color: #cfcfcf; overflow-x: auto; }+.term-body .c { color: #6f6f6f; }+.term-body .w { color: #f2f2f2; }+.term-body .d { color: #9a9a9a; }+.term-body img { margin-top: 14px; border-radius: 10px; width: 100%; aspect-ratio: 16 / 9; object-fit: cover; box-shadow: 0 0 0 1px var(--line); }+.tools { display: flex; flex-wrap: wrap; gap: 8px; margin-top: 26px; }+.tools code { font-family: var(--mono); font-size: 13px; padding: 6px 10px; border-radius: 8px; background: rgba(255,255,255,0.06); box-shadow: inset 0 0 0 1px var(--line); color: var(--text-2); }+.copy { margin-top: 26px; display: flex; align-items: center; gap: 10px; padding: 12px 12px 12px 16px; border-radius: 14px; background: #0f0f0f; box-shadow: inset 0 0 0 1px var(--line); font-family: var(--mono); font-size: 13px; color: var(--text-2); }+.copy code { flex: 1; min-width: 0; overflow-x: auto; white-space: nowrap; scrollbar-width: none; }+.copy button { flex: none; height: 32px; padding: 0 12px; border-radius: 9px; background: rgba(255,255,255,0.08); font-family: var(--font); font-size: 13px; color: var(--text); }+.copy button:hover { background: rgba(255,255,255,0.14); }++/* Requirements */+.req { margin-top: 48px; display: grid; gap: 16px; grid-template-columns: 1fr; }+@media (min-width: 900px) { .req { grid-template-columns: 1fr 1fr; } }+.req .card { min-height: 0; }+.req h3 { display: flex; align-items: center; gap: 12px; }+.req h3 svg { width: 26px; height: 26px; fill: var(--text); }+.rows { margin-top: 18px; }+.row { display: flex; justify-content: space-between; gap: 18px; padding: 14px 0; border-top: 1px solid var(--line); font-size: 15px; }+.row span:first-child { color: var(--text-3); }+.row span:last-child { text-align: right; color: var(--text); }+.beta { font-size: 11px; font-weight: 600; padding: 2px 6px; border-radius: 5px; background: rgba(255,255,255,0.1); color: var(--text-2); margin-left: 6px; vertical-align: 1px; }++/* Privacy */+.privacy { margin-top: 48px; display: grid; gap: 16px; grid-template-columns: 1fr; }+@media (min-width: 900px) { .privacy { grid-template-columns: repeat(3, 1fr); } }+.privacy .card { min-height: 0; }++/* FAQ */+.faq { margin-top: 40px; max-width: 860px; }+details { border-top: 1px solid var(--line); }+details:last-child { border-bottom: 1px solid var(--line); }+summary { list-style: none; cursor: pointer; padding: 22px 0; font-size: 19px; font-weight: 560; letter-spacing: -0.02em; display: flex; justify-content: space-between; gap: 20px; }+summary::-webkit-details-marker { display: none; }+summary::after { content: "+"; font-weight: 400; font-size: 26px; line-height: 1; color: var(--text-3); transition: transform 0.25s; }+details[open] summary::after { transform: rotate(45deg); }+details p { margin: -6px 0 22px; color: var(--text-2); max-width: 44em; }++/* Maker */+.maker { display: inline-flex; align-items: center; gap: 14px; margin-top: 56px; padding: 10px 22px 10px 10px; border-radius: 999px; background: var(--card); box-shadow: inset 0 0 0 1px var(--line); text-align: left; transition: background 0.2s; }+.maker:hover { background: var(--card-hi); }+.maker-photo { width: 52px; height: 52px; border-radius: 50%; object-fit: cover; flex: none; box-shadow: 0 0 0 2px rgba(255,255,255,0.12); }+.maker b { display: block; font-weight: 600; font-size: 15px; letter-spacing: -0.01em; }+.maker small { display: block; color: var(--text-2); font-size: 13.5px; }+footer .me { color: var(--text-2); text-decoration: underline; text-underline-offset: 3px; }+footer .me:hover { color: var(--text); }++/* Final */+.final { padding: 160px 0 120px; text-align: center; position: relative; }+.final::before { content: ""; position: absolute; left: 50%; top: 40px; width: 1100px; height: 600px; transform: translateX(-50%); background: radial-gradient(ellipse at center, rgba(247,246,242,0.08), transparent 60%); pointer-events: none; }+.final .mark { width: 64px; height: 64px; margin: 0 auto 30px; }+footer { border-top: 1px solid var(--line); padding: 36px 0 48px; font-size: 14px; color: var(--text-3); }+footer .wrap { display: flex; flex-wrap: wrap; gap: 18px 30px; align-items: center; justify-content: space-between; }+footer nav { display: flex; gap: 22px; flex-wrap: wrap; }+footer a:hover { color: var(--text); }+.legal { width: 100%; font-size: 12.5px; color: var(--text-3); opacity: 0.8; }++/* Reveal */+.rv { opacity: 0; transform: translateY(28px); transition: opacity 0.9s cubic-bezier(0.16, 1, 0.3, 1), transform 0.9s cubic-bezier(0.16, 1, 0.3, 1); }+.rv.in { opacity: 1; transform: none; }+.rv.d1 { transition-delay: 0.08s; } .rv.d2 { transition-delay: 0.16s; } .rv.d3 { transition-delay: 0.24s; }+@media (prefers-reduced-motion: reduce) { .rv { opacity: 1; transform: none; transition: none; } html { scroll-behavior: auto; } .mask-demo .path { transition: none; stroke-dashoffset: 0; } .mask-demo .after { transition: none; } }++/* Steps + comparison */+.step { display: inline-grid; place-items: center; width: 34px; height: 34px; border-radius: 50%; background: var(--paper); color: var(--ink); font-weight: 650; font-size: 15px; margin-bottom: 18px; }+.table-wrap { margin-top: 44px; overflow-x: auto; border-radius: var(--radius); box-shadow: inset 0 0 0 1px var(--line); background: linear-gradient(180deg, rgba(255,255,255,0.04), rgba(255,255,255,0.012)); }+.cmp-table { width: 100%; border-collapse: collapse; min-width: 720px; font-size: 15px; }+.cmp-table th, .cmp-table td { padding: 16px 20px; text-align: left; border-top: 1px solid var(--line); vertical-align: top; }+.cmp-table thead th { border-top: 0; font-size: 15px; font-weight: 600; color: var(--text-2); }+.cmp-table thead th span { font-weight: 450; font-size: 13px; color: var(--text-3); }+.cmp-table td:first-child { color: var(--text-2); width: 30%; }+.cmp-table td { color: var(--text-3); }+.cmp-table .us { color: var(--text); font-weight: 560; background: rgba(255,255,255,0.035); }+.cmp-table thead .us { color: var(--paper); }++/* Article pages (guides) */+.article { padding: 120px 0 40px; }+.wrap.narrow { width: min(780px, calc(100% - 32px)); }+.crumbs { font-size: 14px; color: var(--text-3); margin-bottom: 22px; }+.crumbs a:hover { color: var(--text); }+.crumbs span { margin: 0 6px; }+.article h1 { font-size: clamp(36px, 5.4vw, 58px); letter-spacing: -0.04em; line-height: 1.06; margin: 0; max-width: none; }+.article h2 { font-size: clamp(26px, 3.2vw, 34px); letter-spacing: -0.03em; line-height: 1.15; margin: 64px 0 0; max-width: none; }+.article h3 { font-size: 18px; letter-spacing: -0.015em; margin: 28px 0 0; font-weight: 600; }+.article p, .article li { color: var(--text-2); font-size: 17px; }+.article p { margin: 16px 0 0; }+.article strong { color: var(--text); font-weight: 600; }+.article a:not(.btn) { color: var(--text); text-decoration: underline; text-underline-offset: 3px; text-decoration-color: rgba(255,255,255,0.3); }+.article a:not(.btn):hover { text-decoration-color: var(--text); }+.article ul, .article ol { margin: 16px 0 0; padding-left: 22px; }+.article li { margin: 8px 0; }+.article .table-wrap { margin-top: 24px; }+.article .cmp-table { min-width: 560px; }+.lede-left { font-size: 20px !important; margin-top: 22px !important; letter-spacing: -0.01em; }+.article figure { margin: 44px 0 0; }+.article figure img { border-radius: 16px; box-shadow: 0 0 0 1px var(--line-hi), 0 40px 120px rgba(0,0,0,0.6); width: 100%; height: auto; }+.article figcaption { margin-top: 12px; font-size: 14px; color: var(--text-3); text-align: center; }+.article .term { margin-top: 40px; }+.article code { font-family: var(--mono); font-size: 0.88em; padding: 2px 6px; border-radius: 6px; background: rgba(255,255,255,0.07); color: var(--text); }+.article pre { margin: 14px 0 0; padding: 16px 18px; border-radius: 14px; background: #0f0f0f; box-shadow: inset 0 0 0 1px var(--line); overflow-x: auto; }+.article pre code { padding: 0; background: none; font-size: 13.5px; line-height: 1.6; color: #d6d6d6; }+.note { font-size: 15px !important; color: var(--text-3) !important; }+.cta-box { margin: 72px 0 40px; padding: 36px; border-radius: var(--radius); background: linear-gradient(180deg, rgba(255,255,255,0.06), rgba(255,255,255,0.02)); box-shadow: inset 0 0 0 1px var(--line); }+.cta-box h2 { margin: 0 !important; }diff --git a/site/index.html b/site/index.htmlindex bfa539e..b8a8a59 100644--- a/site/index.html+++ b/site/index.html@@ -3,227 +3,178 @@<head><meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1">-<title>Kokon — Edit photos by describing the change</title>-<meta name="description" content="Kokon edits and creates photos on your own GPU with Qwen-Image 2.1. One installer for Windows and Mac, no Python, no cloud, no account. Works with your AI agent.">-<meta name="theme-color" content="#0a0a0a">-<link rel="icon" href="assets/favicon.svg" type="image/svg+xml">-<link rel="icon" href="assets/favicon.ico" sizes="any">-<link rel="apple-touch-icon" href="assets/apple-touch-icon.png">-<meta property="og:title" content="Kokon — Edit photos by describing the change">-<meta property="og:description" content="Qwen-Image 2.1 on your own GPU. One installer, no Python, no cloud. Free and open source.">+<title>Kokon: Run Qwen-Image 2.1 Locally — Free AI Photo Editor for Windows & Mac</title>+<meta name="description" content="Free, open-source app to edit and create photos with Qwen-Image 2.1 on your own GPU. No Python, no ComfyUI, no cloud. Windows (NVIDIA, AMD, Intel) and Apple Silicon Macs. MCP server for AI agents.">+<meta name="keywords" content="Qwen-Image 2.1, Qwen Image Edit, run Qwen-Image locally, local AI photo editor, offline AI image editor, ComfyUI alternative, AI inpainting, remove objects from photos, MCP image generation, stable-diffusion.cpp">+<meta name="author" content="Gabriele Lozupone">+<meta name="robots" content="index, follow, max-image-preview:large, max-video-preview:-1">+<meta property="og:site_name" content="Kokon">+<meta property="og:title" content="Kokon — Edit photos by describing the change, on your own GPU">+<meta property="og:description" content="Qwen-Image 2.1 on your own computer. One installer for Windows and Mac, no Python, no cloud. Free and open source."><meta property="og:image" content="https://gabrielelozupone.github.io/kokon/assets/og.png">+<meta property="og:image:width" content="1280">+<meta property="og:image:height" content="640">+<meta property="og:image:alt" content="Kokon logo and tagline"><meta property="og:url" content="https://gabrielelozupone.github.io/kokon/">-<link rel="canonical" href="https://gabrielelozupone.github.io/kokon/"><meta property="og:type" content="website">+<meta property="og:video" content="https://gabrielelozupone.github.io/kokon/assets/kokon.mp4">+<meta property="og:video:type" content="video/mp4">+<meta property="og:video:width" content="1920">+<meta property="og:video:height" content="1080">+<link rel="canonical" href="https://gabrielelozupone.github.io/kokon/"><meta name="twitter:card" content="summary_large_image">+<meta name="twitter:title" content="Kokon — Run Qwen-Image 2.1 locally, no Python">+<meta name="twitter:description" content="Edit photos by describing the change, on your own GPU. Free and open source for Windows and Mac.">+<meta name="twitter:image" content="https://gabrielelozupone.github.io/kokon/assets/og.png">+<link rel="preload" as="image" href="assets/kokon-poster.jpg">+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@graph": [+ {+ "@type": "SoftwareApplication",+ "@id": "https://gabrielelozupone.github.io/kokon/#app",+ "name": "Kokon",+ "alternateName": "Kokon AI photo editor",+ "description": "Free, open-source desktop app that edits and creates photos with Qwen-Image 2.1 on your own GPU. Paint an area to change only that area, compare versions, and let AI agents use it through a built-in MCP server. No Python, no cloud.",+ "url": "https://gabrielelozupone.github.io/kokon/",+ "downloadUrl": "https://github.com/GabrieleLozupone/kokon/releases/latest",+ "installUrl": "https://github.com/GabrieleLozupone/kokon/releases/latest",+ "softwareVersion": "0.9.0",+ "applicationCategory": "MultimediaApplication",+ "applicationSubCategory": "AI image editor",+ "operatingSystem": "Windows 10, Windows 11, macOS 26",+ "processorRequirements": "NVIDIA, AMD or Intel GPU with 12 GB of VRAM, or Apple Silicon",+ "storageRequirements": "15 GB",+ "fileSize": "35 MB",+ "isAccessibleForFree": true,+ "license": "https://www.gnu.org/licenses/agpl-3.0.html",+ "offers": { "@type": "Offer", "price": "0", "priceCurrency": "USD" },+ "screenshot": "https://gabrielelozupone.github.io/kokon/assets/kokon-poster.jpg",+ "image": "https://gabrielelozupone.github.io/kokon/assets/og.png",+ "featureList": ["Edit photos with a sentence", "Create images from text", "Paint an area to change only that area (inpainting)", "Remove, replace and add objects", "Expand images beyond their borders (outpainting)", "Before/after comparison and version history", "Private mode", "MCP server and local HTTP API for AI agents", "Runs offline on NVIDIA, AMD, Intel and Apple Silicon"],+ "author": { "@id": "https://gabrielelozupone.github.io/#me" },+ "sameAs": ["https://github.com/GabrieleLozupone/kokon"]+ },+ {+ "@type": "Person",+ "@id": "https://gabrielelozupone.github.io/#me",+ "name": "Gabriele Lozupone",+ "url": "https://gabrielelozupone.github.io/",+ "jobTitle": "Postdoctoral Research Fellow in AI",+ "sameAs": ["https://github.com/GabrieleLozupone"]+ },+ {+ "@type": "VideoObject",+ "name": "Kokon: edit photos by describing the change, on your own GPU",+ "description": "Kokon in use: creating an image from a sentence, painting an area to remove a scooter, adding a cat, and six more edits made with Qwen-Image 2.1 on a local GPU.",+ "thumbnailUrl": "https://gabrielelozupone.github.io/kokon/assets/kokon-poster.jpg",+ "contentUrl": "https://gabrielelozupone.github.io/kokon/assets/kokon.mp4",+ "uploadDate": "2026-10-10",+ "duration": "PT1M27S"+ },+ {+ "@type": "WebSite",+ "@id": "https://gabrielelozupone.github.io/kokon/#site",+ "name": "Kokon",+ "url": "https://gabrielelozupone.github.io/kokon/",+ "publisher": { "@id": "https://gabrielelozupone.github.io/#me" }+ }+ ]+}+</script><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:opsz,[email protected],400..700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">-<style>-:root {- --bg: #070707;- --bg-2: #0d0d0d;- --card: rgba(255, 255, 255, 0.035);- --card-hi: rgba(255, 255, 255, 0.06);- --line: rgba(255, 255, 255, 0.08);- --line-hi: rgba(255, 255, 255, 0.14);- --text: #f2f1ed;- --text-2: rgba(242, 241, 237, 0.66);- --text-3: rgba(242, 241, 237, 0.42);- --paper: #f7f6f2;- --ink: #0d0d0d;- --radius: 22px;- --max: 1200px;- --font: "Inter", system-ui, -apple-system, "Segoe UI", sans-serif;- --mono: "JetBrains Mono", "Cascadia Code", Consolas, monospace;- color-scheme: dark;+<link rel="stylesheet" href="assets/style.css">+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@type": "FAQPage",+ "mainEntity": [+ {+ "@type": "Question",+ "name": "Is it really free?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Yes. Kokon is open source under the AGPL-3.0 license. No account, no credits, no subscription. It uses your own GPU, so there is nothing to pay for."+ }+ },+ {+ "@type": "Question",+ "name": "How do I run Qwen-Image 2.1 locally without ComfyUI?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Install Kokon and let its setup download the model. Kokon runs Qwen-Image 2.1 through stable-diffusion.cpp, a native engine, so there is no Python, no ComfyUI and no workflow to build. The step-by-step guide is at gabrielelozupone.github.io/kokon/qwen-image-2-1-local."+ }+ },+ {+ "@type": "Question",+ "name": "Which model does it use?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Qwen-Image 2.1 Turbo, the official 8-step version of Qwen's image model, which both creates and edits images and leads the open-weights image-editing rankings."+ }+ },+ {+ "@type": "Question",+ "name": "Does it work on AMD or Intel GPUs?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Yes, in beta, through the Vulkan engine on Windows with 12 GB of VRAM or more. On NVIDIA it uses CUDA; on Apple Silicon Macs it uses Metal."+ }+ },+ {+ "@type": "Question",+ "name": "My GPU has 8 GB. Will it work?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Not well today. 12 GB is the smallest size that gives good results with this model. On a Mac, 16 GB of unified memory works but slowly; 32 GB is recommended."+ }+ },+ {+ "@type": "Question",+ "name": "Can Claude Code or Cursor generate images with it?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Yes. Kokon ships an MCP server: register it once and any MCP client can create, edit, inpaint, expand and upscale images on your GPU, for free and without an API key."+ }+ },+ {+ "@type": "Question",+ "name": "Is it an alternative to Photoshop Generative Fill?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "For many edits, yes: paint an area and describe what should be there, or remove, replace and add objects. It runs offline and costs nothing, but it is an AI editor, not a full layer-based image editor."+ }+ },+ {+ "@type": "Question",+ "name": "Can I use the images commercially?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "The app's code is AGPL-3.0. The model weights have their own licenses: read the Qwen-Image model page before commercial use. The optional upscaler is non-commercial (CC BY-NC-SA 4.0)."+ }+ },+ {+ "@type": "Question",+ "name": "Windows says the app is unrecognized. Is it safe?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "The beta installer is not code-signed yet, so SmartScreen warns about it. Click \"More info\" and then \"Run anyway\". The code is public and every download the app makes is checked against a pinned hash."+ }+ },+ {+ "@type": "Question",+ "name": "Does it work offline?",+ "acceptedAnswer": {+ "@type": "Answer",+ "text": "Yes. After setup has downloaded the models, everything runs offline."+ }+ }+ ]}-* { box-sizing: border-box; }-html { scroll-behavior: smooth; -webkit-text-size-adjust: 100%; }-body { margin: 0; background: var(--bg); color: var(--text); font-family: var(--font); font-size: 17px; line-height: 1.55; -webkit-font-smoothing: antialiased; overflow-x: hidden; }-img, video { display: block; max-width: 100%; }-a { color: inherit; text-decoration: none; }-button { font: inherit; color: inherit; background: none; border: 0; cursor: pointer; padding: 0; }-.wrap { width: min(var(--max), calc(100% - 32px)); margin: 0 auto; }-@media (min-width: 720px) { .wrap { width: min(var(--max), calc(100% - 64px)); } }--/* Grain + light */-body::before { content: ""; position: fixed; inset: 0; pointer-events: none; z-index: 100; opacity: 0.035; background-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='160' height='160'%3E%3Cfilter id='n'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.9' numOctaves='2' stitchTiles='stitch'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' filter='url(%23n)'/%3E%3C/svg%3E"); }--/* Nav */-.nav { position: fixed; top: 0; left: 0; right: 0; z-index: 50; transition: background 0.3s, border-color 0.3s, backdrop-filter 0.3s; border-bottom: 1px solid transparent; }-.nav.scrolled { background: rgba(7, 7, 7, 0.72); backdrop-filter: saturate(1.4) blur(18px); -webkit-backdrop-filter: saturate(1.4) blur(18px); border-color: var(--line); }-.nav .wrap { height: 64px; display: flex; align-items: center; gap: 28px; }-.logo { display: flex; align-items: center; gap: 10px; font-weight: 640; font-size: 19px; letter-spacing: -0.03em; }-.logo svg { width: 22px; height: 22px; }-.nav-links { display: none; gap: 26px; font-size: 14.5px; color: var(--text-2); }-.nav-links a:hover { color: var(--text); }-.nav .spacer { flex: 1; }-.gh { display: none; align-items: center; gap: 8px; font-size: 14px; color: var(--text-2); }-.gh:hover { color: var(--text); }-@media (min-width: 900px) { .nav-links, .gh { display: flex; } }--.btn { display: inline-flex; align-items: center; justify-content: center; gap: 10px; height: 48px; padding: 0 22px; border-radius: 999px; font-weight: 560; font-size: 15.5px; letter-spacing: -0.01em; transition: transform 0.2s, background 0.2s, box-shadow 0.2s; white-space: nowrap; }-.btn svg { width: 18px; height: 18px; flex: none; }-.btn.primary { background: var(--paper); color: var(--ink); box-shadow: 0 0 0 1px rgba(255,255,255,0.2), 0 10px 40px rgba(247,246,242,0.12); }-.btn.primary:hover { transform: translateY(-1px); box-shadow: 0 0 0 1px rgba(255,255,255,0.3), 0 14px 50px rgba(247,246,242,0.2); }-.btn.ghost { background: var(--card-hi); box-shadow: inset 0 0 0 1px var(--line-hi); }-.btn.ghost:hover { background: rgba(255,255,255,0.1); }-.btn.small { height: 38px; padding: 0 16px; font-size: 14px; }-.btn .tag { font-size: 11.5px; font-weight: 600; padding: 2px 7px; border-radius: 6px; background: rgba(255,255,255,0.1); color: var(--text-2); letter-spacing: 0.02em; }--/* Hero */-.hero { position: relative; padding: 148px 0 40px; text-align: center; }-.hero::before { content: ""; position: absolute; left: 50%; top: -240px; width: 1400px; height: 900px; transform: translateX(-50%); background: radial-gradient(ellipse at center, rgba(247,246,242,0.09) 0%, rgba(247,246,242,0) 60%); pointer-events: none; }-.eyebrow { display: inline-flex; align-items: center; gap: 10px; padding: 6px 14px 6px 6px; border-radius: 999px; background: var(--card); box-shadow: inset 0 0 0 1px var(--line); font-size: 13.5px; color: var(--text-2); }-.eyebrow b { font-weight: 600; color: var(--ink); background: var(--paper); padding: 3px 9px; border-radius: 999px; font-size: 12px; }-h1 { font-size: clamp(44px, 8.4vw, 104px); line-height: 1; letter-spacing: -0.045em; font-weight: 680; margin: 26px auto 0; max-width: 11.5em; }-h1 .soft { color: var(--text-3); }-.lede { font-size: clamp(17px, 2vw, 21px); color: var(--text-2); max-width: 36em; margin: 26px auto 0; letter-spacing: -0.012em; }-.ctas { display: flex; flex-wrap: wrap; gap: 12px; justify-content: center; margin-top: 36px; }-.fine { margin-top: 18px; font-size: 13.5px; color: var(--text-3); }--.showreel { position: relative; margin: 72px auto 0; max-width: 1120px; }-.showreel .frame { position: relative; border-radius: 22px; overflow: hidden; background: #000 url("assets/kokon-poster.jpg") center / cover; box-shadow: 0 0 0 1px var(--line-hi), 0 60px 160px rgba(0,0,0,0.75), 0 0 120px rgba(247,246,242,0.06); aspect-ratio: 16 / 9; }-.showreel video { width: 100%; height: 100%; object-fit: cover; }-.showreel .glow { position: absolute; inset: 8% 4% -6%; background: radial-gradient(ellipse at center, rgba(255,214,140,0.16), transparent 65%); filter: blur(40px); z-index: -1; }-.sound { position: absolute; right: 16px; bottom: 16px; height: 36px; padding: 0 14px; border-radius: 999px; background: rgba(0,0,0,0.55); backdrop-filter: blur(10px); font-size: 13px; color: #fff; display: inline-flex; align-items: center; gap: 8px; box-shadow: inset 0 0 0 1px rgba(255,255,255,0.15); }--/* Runs on */-.runs { padding: 64px 0 24px; text-align: center; }-.runs p { font-size: 13px; letter-spacing: 0.16em; text-transform: uppercase; color: var(--text-3); margin: 0 0 22px; }-.runs ul { list-style: none; margin: 0; padding: 0; display: flex; flex-wrap: wrap; justify-content: center; gap: 14px 44px; }-.runs li { display: flex; align-items: center; gap: 10px; color: var(--text-2); font-weight: 560; font-size: 17px; letter-spacing: -0.01em; }-.runs svg { width: 24px; height: 24px; fill: currentColor; }--/* Sections */-section { position: relative; }-.sec { padding: 120px 0 0; }-.kicker { font-size: 13px; letter-spacing: 0.16em; text-transform: uppercase; color: var(--text-3); margin: 0 0 16px; }-h2 { font-size: clamp(36px, 5.6vw, 68px); line-height: 1.04; letter-spacing: -0.042em; font-weight: 670; margin: 0; max-width: 14em; }-.sec-lede { font-size: 19px; color: var(--text-2); max-width: 34em; margin: 20px 0 0; letter-spacing: -0.01em; }-.center { text-align: center; }-.center h2, .center .sec-lede { margin-left: auto; margin-right: auto; }--/* Compare */-.studio { margin-top: 48px; border-radius: 28px; background: linear-gradient(180deg, rgba(255,255,255,0.05), rgba(255,255,255,0.015)); box-shadow: inset 0 0 0 1px var(--line); padding: 14px; }-.tabs { display: flex; gap: 4px; overflow-x: auto; scrollbar-width: none; padding: 4px; background: rgba(0,0,0,0.35); border-radius: 16px; width: max-content; max-width: 100%; margin: 0 auto 14px; }-.tabs::-webkit-scrollbar { display: none; }-.tab { height: 38px; padding: 0 16px; border-radius: 12px; font-size: 14.5px; font-weight: 520; color: var(--text-2); white-space: nowrap; transition: background 0.2s, color 0.2s; }-.tab:hover { color: var(--text); }-.tab.on { background: var(--paper); color: var(--ink); }-.cmp { position: relative; border-radius: 16px; overflow: hidden; background: #000; user-select: none; touch-action: pan-y; cursor: ew-resize; aspect-ratio: 16 / 9; }-.cmp img { position: absolute; inset: 0; width: 100%; height: 100%; object-fit: contain; pointer-events: none; }-.cmp .before { clip-path: inset(0 calc(100% - var(--split, 50%)) 0 0); }-.cmp .handle { position: absolute; top: 0; bottom: 0; left: var(--split, 50%); width: 2px; margin-left: -1px; background: #fff; box-shadow: 0 0 0 1px rgba(0,0,0,0.25), 0 0 24px rgba(255,255,255,0.4); pointer-events: none; }-.cmp .handle span { position: absolute; top: 50%; left: 50%; width: 44px; height: 44px; margin: -22px 0 0 -22px; border-radius: 50%; background: #fff; box-shadow: 0 4px 18px rgba(0,0,0,0.45); display: grid; place-items: center; color: #111; }-.cmp .lbl { position: absolute; top: 14px; padding: 4px 10px; border-radius: 8px; font-size: 12.5px; color: #fff; background: rgba(0,0,0,0.55); backdrop-filter: blur(8px); pointer-events: none; }-.cmp .lbl.l { left: 14px; } .cmp .lbl.r { right: 14px; }-.cmp .bg-blur { position: absolute; inset: -40px; width: calc(100% + 80px); height: calc(100% + 80px); object-fit: cover; filter: blur(40px) brightness(0.35) saturate(1.2); transform: scale(1.1); }-.promptbar { margin-top: 14px; display: flex; align-items: center; gap: 14px; padding: 14px 14px 14px 20px; border-radius: 18px; background: #1a1a1a; box-shadow: inset 0 0 0 1px var(--line); }-.promptbar .mode { flex: none; font-size: 12.5px; font-weight: 560; padding: 4px 10px; border-radius: 8px; background: rgba(255,255,255,0.08); color: var(--text-2); }-.promptbar .txt { flex: 1; min-width: 0; font-size: 16px; color: var(--text); letter-spacing: -0.01em; }-.promptbar .send { flex: none; width: 38px; height: 38px; border-radius: 50%; background: var(--paper); color: var(--ink); display: grid; place-items: center; }-.honest { margin-top: 18px; text-align: center; font-size: 13.5px; color: var(--text-3); }--/* Bento */-.bento { margin-top: 56px; display: grid; gap: 16px; grid-template-columns: 1fr; }-@media (min-width: 900px) { .bento { grid-template-columns: repeat(6, 1fr); } .b-wide { grid-column: span 4; } .b-mid { grid-column: span 3; } .b-small { grid-column: span 2; } }-.card { position: relative; border-radius: var(--radius); background: linear-gradient(180deg, rgba(255,255,255,0.05), rgba(255,255,255,0.018)); box-shadow: inset 0 0 0 1px var(--line); padding: 30px; overflow: hidden; min-height: 280px; display: flex; flex-direction: column; }-.card h3 { margin: 0; font-size: 24px; letter-spacing: -0.035em; font-weight: 640; line-height: 1.15; }-.card p { margin: 10px 0 0; color: var(--text-2); font-size: 15.5px; max-width: 30em; }-.card .vis { margin-top: auto; padding-top: 26px; }-.icon { width: 40px; height: 40px; border-radius: 12px; background: rgba(255,255,255,0.07); display: grid; place-items: center; margin-bottom: 20px; box-shadow: inset 0 0 0 1px var(--line); }-.icon svg { width: 20px; height: 20px; stroke: var(--text); fill: none; stroke-width: 1.8; stroke-linecap: round; stroke-linejoin: round; }-.mask-demo { position: relative; border-radius: 14px; overflow: hidden; aspect-ratio: 16 / 9; }-.mask-demo img { width: 100%; height: 100%; object-fit: cover; }-.mask-demo svg { position: absolute; inset: 0; width: 100%; height: 100%; mix-blend-mode: difference; opacity: 0.75; }-.mask-demo .path { stroke-dasharray: 3000; stroke-dashoffset: 3000; transition: stroke-dashoffset 2.6s cubic-bezier(0.65, 0, 0.35, 1); }-.mask-demo.in .path { stroke-dashoffset: 0; }-.mask-demo .after { position: absolute; inset: 0; opacity: 0; transition: opacity 0.9s 2.8s; }-.mask-demo.in .after { opacity: 1; }-.chips { display: flex; flex-wrap: wrap; gap: 8px; }-.chip { padding: 8px 14px; border-radius: 999px; font-size: 13.5px; background: rgba(255,255,255,0.06); box-shadow: inset 0 0 0 1px var(--line); color: var(--text-2); }-.chip.on { background: var(--paper); color: var(--ink); box-shadow: none; font-weight: 560; }-.timeline { display: flex; align-items: center; gap: 0; }-.timeline .t { width: 64px; height: 64px; border-radius: 14px; overflow: hidden; flex: none; outline: 2px solid transparent; outline-offset: 3px; }-.timeline .t.sel { outline-color: var(--text); }-.timeline .t img { width: 100%; height: 100%; object-fit: cover; }-.timeline .j { width: 22px; height: 2px; background: var(--line-hi); flex: none; }-.timeline .lbls { display: flex; gap: 0; margin-top: 10px; font-size: 12px; color: var(--text-3); }-.big-num { font-size: 64px; font-weight: 680; letter-spacing: -0.05em; line-height: 1; }-.big-num small { font-size: 20px; font-weight: 520; color: var(--text-2); letter-spacing: -0.01em; margin-left: 6px; }--/* Agent */-.split { display: grid; gap: 48px; align-items: center; grid-template-columns: 1fr; }-@media (min-width: 980px) { .split { grid-template-columns: 1fr 1.12fr; gap: 72px; } }-.term { border-radius: 18px; background: #0c0c0c; box-shadow: 0 0 0 1px var(--line-hi), 0 40px 120px rgba(0,0,0,0.6); overflow: hidden; font-family: var(--mono); font-size: 13.5px; line-height: 1.7; }-.term-head { height: 40px; display: flex; align-items: center; gap: 8px; padding: 0 16px; background: #151515; font-family: var(--font); font-size: 12.5px; color: var(--text-3); }-.term-head i { width: 11px; height: 11px; border-radius: 50%; background: #333; }-.term-body { padding: 20px 22px 22px; color: #cfcfcf; overflow-x: auto; }-.term-body .c { color: #6f6f6f; }-.term-body .w { color: #f2f2f2; }-.term-body .d { color: #9a9a9a; }-.term-body img { margin-top: 14px; border-radius: 10px; width: 100%; aspect-ratio: 16 / 9; object-fit: cover; box-shadow: 0 0 0 1px var(--line); }-.tools { display: flex; flex-wrap: wrap; gap: 8px; margin-top: 26px; }-.tools code { font-family: var(--mono); font-size: 13px; padding: 6px 10px; border-radius: 8px; background: rgba(255,255,255,0.06); box-shadow: inset 0 0 0 1px var(--line); color: var(--text-2); }-.copy { margin-top: 26px; display: flex; align-items: center; gap: 10px; padding: 12px 12px 12px 16px; border-radius: 14px; background: #0f0f0f; box-shadow: inset 0 0 0 1px var(--line); font-family: var(--mono); font-size: 13px; color: var(--text-2); }-.copy code { flex: 1; min-width: 0; overflow-x: auto; white-space: nowrap; scrollbar-width: none; }-.copy button { flex: none; height: 32px; padding: 0 12px; border-radius: 9px; background: rgba(255,255,255,0.08); font-family: var(--font); font-size: 13px; color: var(--text); }-.copy button:hover { background: rgba(255,255,255,0.14); }--/* Requirements */-.req { margin-top: 48px; display: grid; gap: 16px; grid-template-columns: 1fr; }-@media (min-width: 900px) { .req { grid-template-columns: 1fr 1fr; } }-.req .card { min-height: 0; }-.req h3 { display: flex; align-items: center; gap: 12px; }-.req h3 svg { width: 26px; height: 26px; fill: var(--text); }-.rows { margin-top: 18px; }-.row { display: flex; justify-content: space-between; gap: 18px; padding: 14px 0; border-top: 1px solid var(--line); font-size: 15px; }-.row span:first-child { color: var(--text-3); }-.row span:last-child { text-align: right; color: var(--text); }-.beta { font-size: 11px; font-weight: 600; padding: 2px 6px; border-radius: 5px; background: rgba(255,255,255,0.1); color: var(--text-2); margin-left: 6px; vertical-align: 1px; }--/* Privacy */-.privacy { margin-top: 48px; display: grid; gap: 16px; grid-template-columns: 1fr; }-@media (min-width: 900px) { .privacy { grid-template-columns: repeat(3, 1fr); } }-.privacy .card { min-height: 0; }--/* FAQ */-.faq { margin-top: 40px; max-width: 860px; }-details { border-top: 1px solid var(--line); }-details:last-child { border-bottom: 1px solid var(--line); }-summary { list-style: none; cursor: pointer; padding: 22px 0; font-size: 19px; font-weight: 560; letter-spacing: -0.02em; display: flex; justify-content: space-between; gap: 20px; }-summary::-webkit-details-marker { display: none; }-summary::after { content: "+"; font-weight: 400; font-size: 26px; line-height: 1; color: var(--text-3); transition: transform 0.25s; }-details[open] summary::after { transform: rotate(45deg); }-details p { margin: -6px 0 22px; color: var(--text-2); max-width: 44em; }--/* Maker */-.maker { display: inline-flex; align-items: center; gap: 14px; margin-top: 56px; padding: 10px 22px 10px 10px; border-radius: 999px; background: var(--card); box-shadow: inset 0 0 0 1px var(--line); text-align: left; transition: background 0.2s; }-.maker:hover { background: var(--card-hi); }-.maker-photo { width: 52px; height: 52px; border-radius: 50%; object-fit: cover; flex: none; box-shadow: 0 0 0 2px rgba(255,255,255,0.12); }-.maker b { display: block; font-weight: 600; font-size: 15px; letter-spacing: -0.01em; }-.maker small { display: block; color: var(--text-2); font-size: 13.5px; }-footer .me { color: var(--text-2); text-decoration: underline; text-underline-offset: 3px; }-footer .me:hover { color: var(--text); }--/* Final */-.final { padding: 160px 0 120px; text-align: center; position: relative; }-.final::before { content: ""; position: absolute; left: 50%; top: 40px; width: 1100px; height: 600px; transform: translateX(-50%); background: radial-gradient(ellipse at center, rgba(247,246,242,0.08), transparent 60%); pointer-events: none; }-.final .mark { width: 64px; height: 64px; margin: 0 auto 30px; }-footer { border-top: 1px solid var(--line); padding: 36px 0 48px; font-size: 14px; color: var(--text-3); }-footer .wrap { display: flex; flex-wrap: wrap; gap: 18px 30px; align-items: center; justify-content: space-between; }-footer nav { display: flex; gap: 22px; flex-wrap: wrap; }-footer a:hover { color: var(--text); }-.legal { width: 100%; font-size: 12.5px; color: var(--text-3); opacity: 0.8; }--/* Reveal */-.rv { opacity: 0; transform: translateY(28px); transition: opacity 0.9s cubic-bezier(0.16, 1, 0.3, 1), transform 0.9s cubic-bezier(0.16, 1, 0.3, 1); }-.rv.in { opacity: 1; transform: none; }-.rv.d1 { transition-delay: 0.08s; } .rv.d2 { transition-delay: 0.16s; } .rv.d3 { transition-delay: 0.24s; }-@media (prefers-reduced-motion: reduce) { .rv { opacity: 1; transform: none; transition: none; } html { scroll-behavior: auto; } .mask-demo .path { transition: none; stroke-dashoffset: 0; } .mask-demo .after { transition: none; } }-</style>+</script></head><body>@@ -246,6 +197,7 @@ footer a:hover { color: var(--text); }<a href="#examples">Examples</a><a href="#features">Features</a><a href="#agents">AI agents</a>+ <a href="qwen-image-2-1-local/">Guide</a><a href="#requirements">Requirements</a><a href="#faq">FAQ</a></nav>@@ -260,7 +212,7 @@ footer a:hover { color: var(--text); }<div class="wrap"><span class="eyebrow rv"><b>Beta</b>Qwen-Image 2.1 Turbo, on your own computer</span><h1 class="rv d1">Edit photos by describing the change.</h1>- <p class="lede rv d2">Kokon runs the best open image-editing model on your own GPU. One installer for Windows and Mac. No Python, no cloud, no account.</p>+ <p class="lede rv d2">Kokon runs Qwen-Image 2.1, the best open image-editing model, on your own GPU. One installer for Windows and Mac. No Python, no ComfyUI, no cloud, no account.</p><div class="ctas rv d3" id="download"><a class="btn primary" href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener"><svg fill="currentColor"><use href="#win"/></svg>Download for Windows</a><a class="btn ghost" href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener"><svg fill="currentColor"><use href="#apple"/></svg>Download for macOS <span class="tag">BETA</span></a>@@ -373,6 +325,42 @@ footer a:hover { color: var(--text); }</div></section>+ <section class="sec" id="how">+ <div class="wrap">+ <p class="kicker rv">How it works</p>+ <h2 class="rv d1">Run Qwen-Image 2.1 locally in three steps.</h2>+ <p class="sec-lede rv d2">No Python environment, no ComfyUI workflows, no model files to hunt down. <a href="qwen-image-2-1-local/" style="text-decoration:underline">Read the full guide</a>.</p>+ <div class="privacy">+ <article class="card rv"><span class="step">1</span><h3>Download the installer</h3><p>A 35 MB native app for Windows 10/11 or macOS on Apple Silicon. It installs for your user, no admin rights needed.</p></article>+ <article class="card rv d1"><span class="step">2</span><h3>Let it set up the model</h3><p>Kokon detects your GPU (NVIDIA, AMD, Intel or Apple Silicon), picks the right preset and downloads Qwen-Image 2.1 Turbo, with resume and SHA-256 checks.</p></article>+ <article class="card rv d2"><span class="step">3</span><h3>Describe the change</h3><p>Drop a photo, write what you want and press Enter. Or paint an area first, and only that area changes.</p></article>+ </div>+ </div>+ </section>++ <section class="sec" id="compare">+ <div class="wrap">+ <p class="kicker rv">Compare</p>+ <h2 class="rv d1">The simple way to run Qwen-Image.</h2>+ <p class="sec-lede rv d2">Node-based tools give you every knob; cloud editors need an account and an upload. Kokon is for people who want the result, on their own machine.</p>+ <div class="table-wrap rv">+ <table class="cmp-table">+ <thead><tr><th></th><th class="us">Kokon</th><th>Node-based tools<br><span>e.g. ComfyUI</span></th><th>Cloud editors</th></tr></thead>+ <tbody>+ <tr><td>Setup</td><td class="us">One installer</td><td>Python environment, custom nodes, workflows</td><td>Account</td></tr>+ <tr><td>Your photos stay on your computer</td><td class="us">Yes</td><td>Yes</td><td>No, they are uploaded</td></tr>+ <tr><td>Cost</td><td class="us">Free</td><td>Free</td><td>Credits or subscription</td></tr>+ <tr><td>Edit with one sentence</td><td class="us">Yes</td><td>After building a workflow</td><td>Yes</td></tr>+ <tr><td>Paint an area, change only that</td><td class="us">Built in</td><td>With an inpainting workflow</td><td>Some</td></tr>+ <tr><td>Settings to tune</td><td class="us">None (Fast, Balanced, Quality)</td><td>Many</td><td>Few</td></tr>+ <tr><td>NVIDIA, AMD, Intel, Apple Silicon</td><td class="us">Yes (AMD, Intel, Mac in beta)</td><td>Depends on the setup</td><td>Any browser</td></tr>+ <tr><td>Local MCP server for AI agents</td><td class="us">Built in, no API key</td><td>Third-party</td><td>Paid API</td></tr>+ </tbody>+ </table>+ </div>+ </div>+ </section>+<section class="sec" id="agents"><div class="wrap split"><div>@@ -380,7 +368,7 @@ footer a:hover { color: var(--text); }<h2 class="rv d1">Your agent can make images now.</h2><p class="sec-lede rv d2">Register Kokon once. Then ask for a hero image, a cleaned-up screenshot or a product shot, and your agent makes it on your GPU. The app doesn't even need to be open.</p><div class="copy rv d3"><code id="mcp-cmd">claude mcp add --scope user kokon -- "$env:LOCALAPPDATA\Kokon\kokon-mcp.exe"</code><button type="button" id="copy-btn">Copy</button></div>- <p class="fine rv d3" style="text-align:left">Windows command shown. Mac and JSON config for other hosts are in the <a href="https://github.com/GabrieleLozupone/kokon#use-it-from-your-ai-agent" style="text-decoration:underline">README</a>.</p>+ <p class="fine rv d3" style="text-align:left">Windows command shown. Mac and other hosts: <a href="mcp-image-generation/" style="text-decoration:underline">MCP setup guide</a>.</p></div><div class="term rv d1" aria-label="An AI agent calling Kokon"><div class="term-head"><i></i><i></i><i></i><span style="margin-left:8px">my-website</span></div>@@ -441,10 +429,14 @@ footer a:hover { color: var(--text); }<h2 class="rv d1">Questions.</h2><div class="faq rv d2"><details><summary>Is it really free?</summary><p>Yes. Kokon is open source under the AGPL-3.0 license. No account, no credits, no subscription. It uses your own GPU, so there is nothing to pay for.</p></details>- <details><summary>Which model does it use?</summary><p>Qwen-Image 2.1 Turbo, the official 8-step version of Qwen's image model, which leads the open-weights image-editing rankings. It runs through stable-diffusion.cpp, so no Python is needed.</p></details>+ <details><summary>How do I run Qwen-Image 2.1 locally without ComfyUI?</summary><p>Install Kokon and let its setup download the model. Kokon runs Qwen-Image 2.1 through stable-diffusion.cpp, a native engine, so there is no Python, no ComfyUI and no workflow to build. The step-by-step guide is at gabrielelozupone.github.io/kokon/qwen-image-2-1-local.</p></details>+ <details><summary>Which model does it use?</summary><p>Qwen-Image 2.1 Turbo, the official 8-step version of Qwen's image model, which both creates and edits images and leads the open-weights image-editing rankings.</p></details>+ <details><summary>Does it work on AMD or Intel GPUs?</summary><p>Yes, in beta, through the Vulkan engine on Windows with 12 GB of VRAM or more. On NVIDIA it uses CUDA; on Apple Silicon Macs it uses Metal.</p></details><details><summary>My GPU has 8 GB. Will it work?</summary><p>Not well today. 12 GB is the smallest size that gives good results with this model. On a Mac, 16 GB of unified memory works but slowly; 32 GB is recommended.</p></details>+ <details><summary>Can Claude Code or Cursor generate images with it?</summary><p>Yes. Kokon ships an MCP server: register it once and any MCP client can create, edit, inpaint, expand and upscale images on your GPU, for free and without an API key.</p></details>+ <details><summary>Is it an alternative to Photoshop Generative Fill?</summary><p>For many edits, yes: paint an area and describe what should be there, or remove, replace and add objects. It runs offline and costs nothing, but it is an AI editor, not a full layer-based image editor.</p></details><details><summary>Can I use the images commercially?</summary><p>The app's code is AGPL-3.0. The model weights have their own licenses: read the Qwen-Image model page before commercial use. The optional upscaler is non-commercial (CC BY-NC-SA 4.0).</p></details>- <details><summary>Windows says the app is unrecognized. Is it safe?</summary><p>The beta installer is not code-signed yet, so SmartScreen warns about it. Click "More info" and then "Run anyway". The code is public and every download the app makes is checked against a pinned hash.</p></details>+ <details><summary>Windows says the app is unrecognized. Is it safe?</summary><p>The beta installer is not code-signed yet, so SmartScreen warns about it. Click "More info" and then "Run anyway". The code is public and every download the app makes is checked against a pinned hash.</p></details><details><summary>Does it work offline?</summary><p>Yes. After setup has downloaded the models, everything runs offline.</p></details></div></div>@@ -474,6 +466,8 @@ footer a:hover { color: var(--text); }<span>Designed and built by <a class="me" href="https://gabrielelozupone.github.io/" rel="author">Gabriele Lozupone</a>, with the help of Claude Code.</span><nav aria-label="Links"><a href="https://github.com/GabrieleLozupone/kokon" rel="noopener">GitHub</a>+ <a href="qwen-image-2-1-local/">Run Qwen-Image locally</a>+ <a href="mcp-image-generation/">MCP image generation</a><a href="https://github.com/GabrieleLozupone/kokon/releases" rel="noopener">Releases</a><a href="https://github.com/GabrieleLozupone/kokon/blob/master/LICENSE" rel="noopener">License</a><a href="https://github.com/GabrieleLozupone/kokon/discussions" rel="noopener">Discussions</a>diff --git a/site/mcp-image-generation/index.html b/site/mcp-image-generation/index.htmlnew file mode 100644index 0000000..7c24c37--- /dev/null+++ b/site/mcp-image-generation/index.html@@ -0,0 +1,176 @@+<!doctype html>+<html lang="en">+<head>+<meta charset="utf-8">+<meta name="viewport" content="width=device-width, initial-scale=1">+<title>Free Local Image Generation MCP Server for Claude Code and Cursor (No API Key)</title>+<meta name="description" content="Let Claude Code, Cursor or any MCP client create and edit images on your own GPU with Qwen-Image 2.1. Free, offline, no API key. Setup for Windows and Mac, tools and example prompts.">+<meta name="author" content="Gabriele Lozupone">+<meta name="robots" content="index, follow, max-image-preview:large">+<meta name="theme-color" content="#0a0a0a">+<link rel="canonical" href="https://gabrielelozupone.github.io/kokon/mcp-image-generation/">+<link rel="icon" href="../assets/favicon.svg" type="image/svg+xml">+<link rel="icon" href="../assets/favicon.ico" sizes="any">+<link rel="apple-touch-icon" href="../assets/apple-touch-icon.png">+<meta property="og:site_name" content="Kokon">+<meta property="og:type" content="article">+<meta property="og:title" content="Free local image generation for Claude Code and MCP clients">+<meta property="og:description" content="An MCP server that creates and edits images on your GPU with Qwen-Image 2.1. No API key, no cloud, no cost.">+<meta property="og:url" content="https://gabrielelozupone.github.io/kokon/mcp-image-generation/">+<meta property="og:image" content="https://gabrielelozupone.github.io/kokon/assets/img/hero.webp">+<meta name="twitter:card" content="summary_large_image">+<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:opsz,[email protected],400..700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">+<link rel="stylesheet" href="../assets/style.css">+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@graph": [+ {+ "@type": "TechArticle",+ "headline": "Free Local Image Generation MCP Server for Claude Code and Cursor (No API Key)",+ "description": "Set up Kokon's MCP server so AI agents can create and edit images on your own GPU with Qwen-Image 2.1.",+ "image": "https://gabrielelozupone.github.io/kokon/assets/img/hero.webp",+ "datePublished": "2026-10-10",+ "dateModified": "2026-10-10",+ "author": { "@type": "Person", "name": "Gabriele Lozupone", "url": "https://gabrielelozupone.github.io/" },+ "about": [{ "@type": "Thing", "name": "Model Context Protocol" }, { "@type": "SoftwareApplication", "name": "Kokon", "url": "https://gabrielelozupone.github.io/kokon/" }],+ "mainEntityOfPage": "https://gabrielelozupone.github.io/kokon/mcp-image-generation/"+ },+ {+ "@type": "BreadcrumbList",+ "itemListElement": [+ { "@type": "ListItem", "position": 1, "name": "Kokon", "item": "https://gabrielelozupone.github.io/kokon/" },+ { "@type": "ListItem", "position": 2, "name": "MCP image generation", "item": "https://gabrielelozupone.github.io/kokon/mcp-image-generation/" }+ ]+ },+ {+ "@type": "FAQPage",+ "mainEntity": [+ { "@type": "Question", "name": "Can Claude Code generate images without an API key?", "acceptedAnswer": { "@type": "Answer", "text": "Yes, with a local MCP server such as Kokon's kokon-mcp: images are made by Qwen-Image 2.1 on your own GPU, so there is no API key and no per-image cost." } },+ { "@type": "Question", "name": "Does the Kokon app need to be open for the MCP server to work?", "acceptedAnswer": { "@type": "Answer", "text": "No. The MCP server starts the engine on demand. The first call takes about a minute while the model loads; without the desktop window open, the model is unloaded after 15 idle minutes." } },+ { "@type": "Question", "name": "Are my images sent to the AI model?", "acceptedAnswer": { "@type": "Answer", "text": "No. Results stay on disk and the tool returns their path. A small JPEG preview is included only if the agent asks for it with include_preview." } }+ ]+ }+ ]+}+</script>+</head>+<body>+<svg width="0" height="0" style="position:absolute" aria-hidden="true">+ <defs>+ <symbol id="mark" viewBox="0 0 1024 1024"><path d="M0.00 242.81 C0.00 112.17 70.23 25.08 175.59 25.08 C280.94 25.08 351.17 112.17 351.17 242.81 L351.17 781.19 C351.17 911.83 280.94 998.92 175.59 998.92 C70.23 998.92 0.00 911.83 0.00 781.19 Z M389.53 548.95 C389.53 250.35 662.35 25.08 1024.00 25.08 C1024.00 323.68 751.18 548.95 389.53 548.95 Z M389.53 587.31 C389.53 821.93 618.71 998.92 922.49 998.92 C922.49 764.30 693.31 587.31 389.53 587.31 Z" fill="currentColor"/></symbol>+ </defs>+</svg>++<header class="nav scrolled" id="nav">+ <div class="wrap">+ <a class="logo" href="../" aria-label="Kokon home"><svg><use href="#mark"/></svg>Kokon</a>+ <nav class="nav-links" aria-label="Sections">+ <a href="../#examples">Examples</a>+ <a href="../#features">Features</a>+ <a href="../qwen-image-2-1-local/">Guide</a>+ <a href="../#faq">FAQ</a>+ </nav>+ <span class="spacer"></span>+ <a class="btn primary small" href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener">Download</a>+ </div>+</header>++<main class="article">+ <div class="wrap narrow">+ <nav class="crumbs" aria-label="Breadcrumb"><a href="../">Kokon</a> <span>/</span> MCP</nav>+ <h1>Free local image generation for Claude Code and MCP clients</h1>+ <p class="lede-left">Most image MCP servers forward your prompt to a paid cloud API. Kokon's runs Qwen-Image 2.1 on your own GPU: no API key, no cost per image, nothing uploaded.</p>++ <div class="term" aria-label="Claude Code calling Kokon">+ <div class="term-head"><i></i><i></i><i></i><span style="margin-left:8px">my-website</span></div>+ <div class="term-body">+ <div class="w"><span class="c">></span> Make a hero image for my landing page: a blue morpho butterfly on a silk cocoon, 16:9</div>+ <div style="margin-top:10px"><span class="w">● kokon</span> <span class="d">· generate_image</span></div>+ <div class="d" style="padding-left:20px">aspect_ratio: "16:9" · quality: "quality"</div>+ <div style="padding-left:20px"><span class="c">⎿</span> Saved public/hero.png · 2752×1536 · 39 s</div>+ <img src="../assets/img/hero.webp" width="1800" height="1005" alt="A blue morpho butterfly on a silk cocoon, generated locally by Kokon for an AI agent">+ </div>+ </div>++ <h2>What your agent can do</h2>+ <div class="table-wrap">+ <table class="cmp-table">+ <thead><tr><th>Tool</th><th class="us">What it does</th></tr></thead>+ <tbody>+ <tr><td><code>generate_image</code></td><td class="us">Create an image from a prompt; aspect ratio from the prompt ("16:9 hero") or explicit</td></tr>+ <tr><td><code>edit_image</code></td><td class="us">Edit a photo with an instruction, keeping its size</td></tr>+ <tr><td><code>edit_images</code></td><td class="us">Combine several images ("put the product from image 1 on the table in image 2")</td></tr>+ <tr><td><code>inpaint_image</code></td><td class="us">Change only a masked area: edit, remove, replace or add</td></tr>+ <tr><td><code>outpaint_image</code></td><td class="us">Extend an image beyond its borders, in pixels or percent</td></tr>+ <tr><td><code>upscale_image</code></td><td class="us">Enlarge up to 8192 px</td></tr>+ <tr><td><code>get_image_info</code></td><td class="us">Size, format and the generation metadata of an image</td></tr>+ <tr><td><code>get_generator_status</code></td><td class="us">GPU, VRAM, whether the model is loaded, the queue</td></tr>+ </tbody>+ </table>+ </div>++ <h2>Set it up</h2>+ <p>Install <a href="../">Kokon</a> and run its first-time setup once (it downloads the model). Then register the MCP server.</p>+ <h3>Claude Code on Windows</h3>+ <pre><code>claude mcp add --scope user kokon -- "$env:LOCALAPPDATA\Kokon\kokon-mcp.exe"</code></pre>+ <h3>Claude Code on Mac</h3>+ <pre><code>claude mcp add --scope user kokon -- /Applications/Kokon.app/Contents/MacOS/kokon-mcp</code></pre>+ <h3>Cursor and other MCP hosts</h3>+ <p>Any host that takes the standard JSON config (on a Mac the command is <code>/Applications/Kokon.app/Contents/MacOS/kokon-mcp</code>):</p>+ <pre><code>{+ "mcpServers": {+ "kokon": {+ "type": "stdio",+ "command": "C:\\Users\\<you>\\AppData\\Local\\Kokon\\kokon-mcp.exe",+ "args": []+ }+ }+}</code></pre>+ <p>The desktop app does not need to be open: the server starts the engine on demand. The first call takes about a minute while the model loads.</p>++ <h2>Prompts to try</h2>+ <ul>+ <li>"Generate a 16:9 hero image for this README and save it in docs/"</li>+ <li>"Remove the background clutter from screenshot.png"</li>+ <li>"Make three product shots of assets/mug.png: marble counter, wooden desk, outdoor café"</li>+ <li>"Extend banner.png to 21:9, continuing the sky"</li>+ <li>"Create an app icon, a flat orange fox on a dark background, 1:1"</li>+ </ul>++ <h2>Safe by default</h2>+ <ul>+ <li>Reads and writes are limited to the project folder, Kokon's output folder and folders you allow with <code>QLS_MCP_ALLOWED_DIRS</code>. Existing files are never overwritten.</li>+ <li>Images stay on disk; the tool returns their path, size, seed and metadata. A small preview goes back to the model only when it asks for one (<code>include_preview</code>).</li>+ <li>Everything runs on <code>127.0.0.1</code>. No account, no telemetry, no prompts in the logs.</li>+ </ul>++ <h2>Requirements</h2>+ <p>A Windows PC with a 12 GB NVIDIA, AMD or Intel Arc GPU, or an Apple Silicon Mac (32 GB recommended). On an RTX 5070 Ti a new image takes 22 to 46 s and an edit 25 to 55 s. Details in the <a href="../qwen-image-2-1-local/">setup guide</a>.</p>++ <div class="cta-box">+ <h2>Give your agent a GPU</h2>+ <p>Kokon is free and open source.</p>+ <div class="ctas" style="justify-content:flex-start">+ <a class="btn primary" href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener">Download Kokon</a>+ <a class="btn ghost" href="https://github.com/GabrieleLozupone/kokon#use-it-from-your-ai-agent" rel="noopener">MCP docs on GitHub</a>+ </div>+ </div>+ </div>+</main>++<footer>+ <div class="wrap">+ <a class="logo" href="../"><svg><use href="#mark"/></svg>Kokon</a>+ <span>Designed and built by <a class="me" href="https://gabrielelozupone.github.io/" rel="author">Gabriele Lozupone</a>, with the help of Claude Code.</span>+ <nav aria-label="Links">+ <a href="../">Home</a>+ <a href="../qwen-image-2-1-local/">Setup guide</a>+ <a href="https://github.com/GabrieleLozupone/kokon" rel="noopener">GitHub</a>+ </nav>+ </div>+</footer>+</body>+</html>diff --git a/site/qwen-image-2-1-local/index.html b/site/qwen-image-2-1-local/index.htmlnew file mode 100644index 0000000..cb2abba--- /dev/null+++ b/site/qwen-image-2-1-local/index.html@@ -0,0 +1,192 @@+<!doctype html>+<html lang="en">+<head>+<meta charset="utf-8">+<meta name="viewport" content="width=device-width, initial-scale=1">+<title>How to Run Qwen-Image 2.1 Locally on Windows and Mac (No Python, No ComfyUI)</title>+<meta name="description" content="Step-by-step guide to run Qwen-Image 2.1 on your own GPU: requirements (12 GB VRAM, AMD, Intel, Apple Silicon), one-installer setup with Kokon, editing tips, speed and troubleshooting.">+<meta name="author" content="Gabriele Lozupone">+<meta name="robots" content="index, follow, max-image-preview:large">+<meta name="theme-color" content="#0a0a0a">+<link rel="canonical" href="https://gabrielelozupone.github.io/kokon/qwen-image-2-1-local/">+<link rel="icon" href="../assets/favicon.svg" type="image/svg+xml">+<link rel="icon" href="../assets/favicon.ico" sizes="any">+<link rel="apple-touch-icon" href="../assets/apple-touch-icon.png">+<meta property="og:site_name" content="Kokon">+<meta property="og:type" content="article">+<meta property="og:title" content="How to run Qwen-Image 2.1 locally (no Python, no ComfyUI)">+<meta property="og:description" content="Run Qwen-Image 2.1 on Windows or Mac with one installer: requirements, setup, editing tips and troubleshooting.">+<meta property="og:url" content="https://gabrielelozupone.github.io/kokon/qwen-image-2-1-local/">+<meta property="og:image" content="https://gabrielelozupone.github.io/kokon/assets/img/app-screenshot.webp">+<meta name="twitter:card" content="summary_large_image">+<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:opsz,[email protected],400..700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">+<link rel="stylesheet" href="../assets/style.css">+<script type="application/ld+json">+{+ "@context": "https://schema.org",+ "@graph": [+ {+ "@type": "TechArticle",+ "headline": "How to Run Qwen-Image 2.1 Locally on Windows and Mac (No Python, No ComfyUI)",+ "description": "Requirements, one-installer setup, editing tips, speed and troubleshooting for running Qwen-Image 2.1 on your own GPU.",+ "image": "https://gabrielelozupone.github.io/kokon/assets/img/app-screenshot.webp",+ "datePublished": "2026-10-10",+ "dateModified": "2026-10-10",+ "author": { "@type": "Person", "name": "Gabriele Lozupone", "url": "https://gabrielelozupone.github.io/" },+ "about": [{ "@type": "Thing", "name": "Qwen-Image 2.1" }, { "@type": "SoftwareApplication", "name": "Kokon", "url": "https://gabrielelozupone.github.io/kokon/" }],+ "mainEntityOfPage": "https://gabrielelozupone.github.io/kokon/qwen-image-2-1-local/"+ },+ {+ "@type": "BreadcrumbList",+ "itemListElement": [+ { "@type": "ListItem", "position": 1, "name": "Kokon", "item": "https://gabrielelozupone.github.io/kokon/" },+ { "@type": "ListItem", "position": 2, "name": "Run Qwen-Image 2.1 locally", "item": "https://gabrielelozupone.github.io/kokon/qwen-image-2-1-local/" }+ ]+ },+ {+ "@type": "FAQPage",+ "mainEntity": [+ { "@type": "Question", "name": "How much VRAM does Qwen-Image 2.1 need?", "acceptedAnswer": { "@type": "Answer", "text": "With Kokon, 12 GB of VRAM is enough (tested on an RTX 5070 12 GB). 16 GB or more is more comfortable. Apple Silicon Macs work with 16 GB of unified memory, slowly; 32 GB is recommended." } },+ { "@type": "Question", "name": "Can I run Qwen-Image 2.1 without ComfyUI or Python?", "acceptedAnswer": { "@type": "Answer", "text": "Yes. Kokon runs it through stable-diffusion.cpp, a native C++ engine, and ships as one installer. It downloads and checks the model on first run." } },+ { "@type": "Question", "name": "Does Qwen-Image 2.1 run on AMD or Intel GPUs?", "acceptedAnswer": { "@type": "Answer", "text": "Yes, through Vulkan on Windows (beta in Kokon), with 12 GB of VRAM or more." } },+ { "@type": "Question", "name": "How fast is Qwen-Image 2.1 locally?", "acceptedAnswer": { "@type": "Answer", "text": "With the official 8-step Turbo model on an RTX 5070 Ti: about 22 s for a 1.5K image, about 40 s for native 2K, and 25 to 55 s for an edit. An M2 Pro Mac takes about 4 minutes for a 1-megapixel image." } }+ ]+ }+ ]+}+</script>+</head>+<body>+<svg width="0" height="0" style="position:absolute" aria-hidden="true">+ <defs>+ <symbol id="mark" viewBox="0 0 1024 1024"><path d="M0.00 242.81 C0.00 112.17 70.23 25.08 175.59 25.08 C280.94 25.08 351.17 112.17 351.17 242.81 L351.17 781.19 C351.17 911.83 280.94 998.92 175.59 998.92 C70.23 998.92 0.00 911.83 0.00 781.19 Z M389.53 548.95 C389.53 250.35 662.35 25.08 1024.00 25.08 C1024.00 323.68 751.18 548.95 389.53 548.95 Z M389.53 587.31 C389.53 821.93 618.71 998.92 922.49 998.92 C922.49 764.30 693.31 587.31 389.53 587.31 Z" fill="currentColor"/></symbol>+ </defs>+</svg>++<header class="nav scrolled" id="nav">+ <div class="wrap">+ <a class="logo" href="../" aria-label="Kokon home"><svg><use href="#mark"/></svg>Kokon</a>+ <nav class="nav-links" aria-label="Sections">+ <a href="../#examples">Examples</a>+ <a href="../#features">Features</a>+ <a href="../mcp-image-generation/">AI agents</a>+ <a href="../#faq">FAQ</a>+ </nav>+ <span class="spacer"></span>+ <a class="btn primary small" href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener">Download</a>+ </div>+</header>++<main class="article">+ <div class="wrap narrow">+ <nav class="crumbs" aria-label="Breadcrumb"><a href="../">Kokon</a> <span>/</span> Guide</nav>+ <h1>How to run Qwen-Image 2.1 locally on Windows and Mac</h1>+ <p class="lede-left">No Python, no ComfyUI, no cloud. One installer, a model download, and you edit photos by describing the change. Updated October 10, 2026.</p>++ <figure>+ <img src="../assets/img/app-screenshot.webp" width="1600" height="980" alt="Kokon running Qwen-Image 2.1 locally: a café street photo with the scooter removed, shown with a before/after slider" loading="eager">+ <figcaption>Kokon with a local Qwen-Image 2.1 edit: the scooter was painted and removed.</figcaption>+ </figure>++ <h2>What Qwen-Image 2.1 is</h2>+ <p>Qwen-Image 2.1 is the open-weights image model from Alibaba's Qwen team. One model both <strong>creates images from text</strong> and <strong>edits existing photos from an instruction</strong> ("remove the people", "make it a snowy night", "change the sign to say…"). It leads the open-weights image-editing rankings, and the official <strong>Turbo</strong> version makes an image in 8 steps, fast enough to use on a consumer GPU.</p>+ <p>Most guides run it through ComfyUI: install Python, add custom nodes, download several model files and load a workflow. That works, but it is a lot to set up and keep working. This guide uses <a href="../">Kokon</a>, a free, open-source desktop app that does all of that for you.</p>++ <h2>What you need</h2>+ <div class="table-wrap">+ <table class="cmp-table">+ <thead><tr><th></th><th class="us">Minimum</th><th>Recommended</th></tr></thead>+ <tbody>+ <tr><td>Windows + NVIDIA</td><td class="us">12 GB VRAM (RTX 3060 12 GB, 4070, 5070)</td><td>16 GB or more (RTX 4080, 5070 Ti, 4090, 5090)</td></tr>+ <tr><td>Windows + AMD / Intel Arc</td><td class="us">12 GB VRAM, Vulkan (beta)</td><td>16 GB or more</td></tr>+ <tr><td>Mac</td><td class="us">Apple Silicon, 16 GB unified memory, macOS 26 (slow)</td><td>32 GB or more, Pro/Max/Ultra chip</td></tr>+ <tr><td>Disk</td><td class="us">15 GB free</td><td>25 GB free</td></tr>+ </tbody>+ </table>+ </div>+ <p class="note">8 GB cards are not supported well: the model is large, and 12 GB is the smallest size that gives good results.</p>++ <h2>Step 1: Install Kokon</h2>+ <p>Download the latest release from <a href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener">GitHub</a>:</p>+ <ul>+ <li><strong>Windows:</strong> <code>Kokon_<version>_x64-setup.exe</code> (about 35 MB). It installs for your user only, no administrator rights. The beta is not code-signed yet, so if SmartScreen warns, click <em>More info → Run anyway</em>.</li>+ <li><strong>Mac (Apple Silicon, beta):</strong> <code>Kokon_<version>_aarch64.dmg</code>. Drag Kokon to Applications. The first time, open <em>System Settings → Privacy & Security</em> and click <em>Open Anyway</em>.</li>+ </ul>++ <h2>Step 2: Let it download the model</h2>+ <p>On first launch Kokon shows your graphics card (or your Mac's chip and memory) and recommends a preset. It downloads what it needs: the stable-diffusion.cpp engine, <strong>Qwen-Image 2.1 Turbo</strong> as a 4-bit GGUF (about 4 GB), the Qwen3-VL text encoder and the VAE, about 12 GB in total. Downloads resume if interrupted, and every file is checked against a pinned SHA-256 hash.</p>+ <p>On Windows with an NVIDIA card you can pick the <strong>CUDA</strong> engine (fastest) or <strong>Vulkan</strong>; AMD and Intel cards use Vulkan. Macs use Metal.</p>++ <h2>Step 3: Create or edit</h2>+ <ol>+ <li><strong>Create:</strong> type a description and press Enter. Add an aspect ratio in words ("16:9 banner") or pick one from the menu.</li>+ <li><strong>Edit:</strong> drop or paste a photo, describe the change, press Enter. The result appears as a new version; drag the slider to compare before and after.</li>+ <li><strong>Edit only one area:</strong> click <em>Paint area</em>, brush over the part to change, then choose <em>Change</em>, <em>Remove</em>, <em>Replace</em> or <em>Add</em>. Pixels outside the brush stay exactly as they were.</li>+ <li><strong>Expand:</strong> click <em>Expand</em> to extend the photo beyond its borders, to any ratio.</li>+ </ol>+ <p>There are only three quality levels: <strong>Fast</strong> (~1 MP), <strong>Balanced</strong> (~1.5K) and <strong>Quality</strong> (native 2K). Resolution, memory use and upscaling are chosen for your GPU.</p>++ <h2>Prompts that work well</h2>+ <div class="table-wrap">+ <table class="cmp-table">+ <thead><tr><th>Edit</th><th class="us">Example prompt</th></tr></thead>+ <tbody>+ <tr><td>Remove</td><td class="us">"Remove all the people from the beach"</td></tr>+ <tr><td>Replace</td><td class="us">"Replace the yellow armchair with a deep green velvet sofa"</td></tr>+ <tr><td>Relight</td><td class="us">"Make it a snowy winter night, warm light glowing from the windows"</td></tr>+ <tr><td>Text</td><td class="us">"Change the gold sign text to 'KOKON CAFÉ' in the same lettering"</td></tr>+ <tr><td>Restyle</td><td class="us">"Turn it into a hand-painted watercolor illustration"</td></tr>+ <tr><td>Product shot</td><td class="us">"Place the sneaker on wet black stone with a soft orange rim light"</td></tr>+ </tbody>+ </table>+ </div>+ <p>Say what should change and, when it matters, what should stay: "…keep the person and the composition the same". For small objects, paint the area first: it is more reliable than describing where the object is.</p>++ <h2>How fast is it?</h2>+ <p>Measured with Qwen-Image 2.1 Turbo (8 steps):</p>+ <ul>+ <li><strong>RTX 5070 Ti (16 GB):</strong> about 22 s for a new 1.5K image, about 40 s at native 2K, 25 to 55 s for an edit.</li>+ <li><strong>RTX 5070 (12 GB):</strong> works with the 12 GB preset, a little slower than the 5070 Ti.</li>+ <li><strong>M2 Pro, 16 GB:</strong> about 4 minutes for a 1-megapixel image.</li>+ </ul>+ <p>Kokon measures your machine with the first image and shows realistic time estimates from then on.</p>++ <h2>Troubleshooting</h2>+ <ul>+ <li><strong>"Out of memory":</strong> choose Fast or Balanced, close other GPU-heavy apps, or run setup again and pick a lighter preset.</li>+ <li><strong>The first image is slow:</strong> the model loads once (and Vulkan compiles its GPU programs on first use). Later images are faster.</li>+ <li><strong>Something else:</strong> Settings → Diagnostics → Copy, then <a href="https://github.com/GabrieleLozupone/kokon/issues" rel="noopener">open an issue</a> with the report.</li>+ </ul>++ <h2>Use it from an AI agent</h2>+ <p>Kokon includes an MCP server, so Claude Code, Cursor or any MCP client can create and edit images on your GPU without an API key. See <a href="../mcp-image-generation/">free local image generation for Claude Code and MCP</a>.</p>++ <h2>License notes</h2>+ <p>Kokon's code is AGPL-3.0. The model weights have their own licenses; read the Qwen-Image model page before commercial use. Kokon is an independent project, not affiliated with Alibaba.</p>++ <div class="cta-box">+ <h2>Try it</h2>+ <p>Free and open source for Windows and macOS.</p>+ <div class="ctas" style="justify-content:flex-start">+ <a class="btn primary" href="https://github.com/GabrieleLozupone/kokon/releases/latest" rel="noopener">Download Kokon</a>+ <a class="btn ghost" href="https://github.com/GabrieleLozupone/kokon" rel="noopener">View on GitHub</a>+ </div>+ </div>+ </div>+</main>++<footer>+ <div class="wrap">+ <a class="logo" href="../"><svg><use href="#mark"/></svg>Kokon</a>+ <span>Designed and built by <a class="me" href="https://gabrielelozupone.github.io/" rel="author">Gabriele Lozupone</a>, with the help of Claude Code.</span>+ <nav aria-label="Links">+ <a href="../">Home</a>+ <a href="../mcp-image-generation/">MCP guide</a>+ <a href="https://github.com/GabrieleLozupone/kokon" rel="noopener">GitHub</a>+ </nav>+ </div>+</footer>+</body>+</html>diff --git a/site/sitemap.xml b/site/sitemap.xmlnew file mode 100644index 0000000..0405262--- /dev/null+++ b/site/sitemap.xml@@ -0,0 +1,16 @@+<?xml version="1.0" encoding="UTF-8"?>+<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9" xmlns:video="http://www.google.com/schemas/sitemap-video/1.1">+ <url>+ <loc>https://gabrielelozupone.github.io/kokon/</loc>+ <lastmod>2026-10-10</lastmod>+ <video:video>+ <video:thumbnail_loc>https://gabrielelozupone.github.io/kokon/assets/kokon-poster.jpg</video:thumbnail_loc>+ <video:title>Kokon: edit photos by describing the change, on your own GPU</video:title>+ <video:description>Kokon in use: creating an image from a sentence, painting an area to remove a scooter, adding a cat, and six more edits made with Qwen-Image 2.1 on a local GPU.</video:description>+ <video:content_loc>https://gabrielelozupone.github.io/kokon/assets/kokon.mp4</video:content_loc>+ <video:duration>87</video:duration>+ </video:video>+ </url>+ <url><loc>https://gabrielelozupone.github.io/kokon/qwen-image-2-1-local/</loc><lastmod>2026-10-10</lastmod></url>+ <url><loc>https://gabrielelozupone.github.io/kokon/mcp-image-generation/</loc><lastmod>2026-10-10</lastmod></url>+</urlset>