Repository Public
1<!doctype html>2<html lang="en">3<head>4 <meta charset="utf-8">5 <meta name="viewport" content="width=device-width, initial-scale=1, viewport-fit=cover">6 <title>MacLustr AI — Intelligence, running locally.</title>7 <meta name="description" content="High-performance AI inference powered by Apple Silicon and MLX. OpenAI-compatible API with streaming, tool calling and validated JSON.">8 <meta name="theme-color" content="#07090c" media="(prefers-color-scheme: dark)">9 <meta name="theme-color" content="#f6f7f9" media="(prefers-color-scheme: light)">10 <link rel="icon" href="/assets/favicon.svg" type="image/svg+xml">11 <script src="/assets/site/theme.js"></script>12 <link rel="stylesheet" href="/assets/tokens.css">13 <link rel="stylesheet" href="/assets/site.css">14 <link rel="stylesheet" href="/assets/site/landing.css">15 <link rel="modulepreload" href="/assets/site/core.js">16 <link rel="modulepreload" href="/assets/site/landing.js">17</head>18<body>19 <a class="skip-link" href="#main">Skip to content</a>20 <header class="site-header">21 <div class="wrap">22 <a class="brand" href="/" aria-label="MacLustr AI — home">23 <img src="/assets/logo.svg" alt="" width="28" height="28">24 <span>MacLustr <span>AI</span></span>25 </a>26 <nav class="site-nav" id="site-nav" aria-label="Primary">27 <a href="/docs" data-nav="docs">Docs</a>28 <a href="/playground" data-nav="playground">Playground</a>29 <a href="/status" data-nav="status">Status</a>30 <a href="/changelog" data-nav="changelog">Changelog</a>31 </nav>32 <div class="header-tools">33 <button type="button" class="icon-btn theme-toggle" data-theme-toggle aria-label="Switch theme">34 <svg class="i-moon" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M20 14.5A8 8 0 0 1 9.5 4a8 8 0 1 0 10.5 10.5Z"/></svg>35 <svg class="i-sun" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" aria-hidden="true"><circle cx="12" cy="12" r="4"/><path d="M12 2.5v2M12 19.5v2M2.5 12h2M19.5 12h2M5.3 5.3l1.4 1.4M17.3 17.3l1.4 1.4M5.3 18.7l1.4-1.4M17.3 6.7l1.4-1.4"/></svg>36 </button>37 <button type="button" class="icon-btn menu-btn" data-menu-btn aria-controls="site-nav" aria-expanded="false" aria-label="Open menu">38 <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" aria-hidden="true"><path d="M4 7h16M4 12h16M4 17h16"/></svg>39 </button>40 </div>41 </div>42 </header>4344 <main id="main" tabindex="-1">45 <section class="hero" data-hero>46 <div class="hero-bg" aria-hidden="true">47 <div class="aurora a1"></div><div class="aurora a2"></div><div class="aurora a3"></div>48 <div class="hero-grid"></div>49 </div>50 <div class="wrap hero-inner">51 <div class="hero-copy">52 <a class="pill" href="/status" data-status-pill data-state="loading" aria-live="polite"><span class="dot" aria-hidden="true"></span><span data-status-text>Checking status…</span></a>53 <h1>MacLustr AI <span class="hero-dash">—</span> <span class="aurora-text">Intelligence, running locally.</span></h1>54 <p class="hero-sub">High-performance AI inference powered by Apple Silicon and MLX.</p>55 <div class="hero-cta">56 <a class="btn btn-primary" href="/docs/getting-started">Get started <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>57 <a class="btn" href="/playground">Open the Playground</a>58 </div>59 <ul class="hero-tags" aria-label="Capabilities">60 <li>OpenAI-compatible</li><li>SSE streaming</li><li>Tool calling</li><li>Validated JSON</li><li>On-device inference</li>61 </ul>62 </div>63 <aside class="model-card" aria-labelledby="model-card-title">64 <div class="model-card-head">65 <span class="eyebrow" id="model-card-title">Active model</span>66 <span class="badge" data-model-available>—</span>67 </div>68 <p class="model-name" data-model-name><span class="skeleton" style="width:14ch"></span></p>69 <p class="model-id"><code data-model-id>—</code></p>70 <dl class="model-facts">71 <div><dt>Context window</dt><dd class="tabular" data-model-context>—</dd></div>72 <div><dt>Quantization</dt><dd data-model-quant>—</dd></div>73 <div><dt>Engine</dt><dd data-model-engine>—</dd></div>74 <div><dt>API version</dt><dd class="tabular" data-api-version>—</dd></div>75 </dl>76 <p class="model-foot" data-model-foot>Live from <code>/api/status</code></p>77 </aside>78 </div>79 </section>8081 <section class="section wrap example" aria-labelledby="example-title">82 <div class="example-copy">83 <span class="eyebrow">One request away</span>84 <h2 id="example-title">If you speak OpenAI, you already speak MacLustr.</h2>85 <p>Keep your SDK, your message format and your streaming loop. Change the base URL, use a MacLustr key, and call the model served on our hardware.</p>86 <dl class="kv">87 <div><dt>Base URL</dt><dd><code>https://llm.spboucher.ai/v1</code></dd></div>88 <div><dt>Model</dt><dd><code>maclustr-qwen3</code></dd></div>89 <div><dt>Auth</dt><dd><code>Authorization: Bearer mlx_sk_…</code></dd></div>90 </dl>91 <div class="seg" role="group" aria-label="Example variant">92 <button type="button" class="seg-btn" data-variant="chat" aria-pressed="true">Completion</button>93 <button type="button" class="seg-btn" data-variant="stream" aria-pressed="false">Streaming</button>94 <button type="button" class="seg-btn" data-variant="tools" aria-pressed="false">Tool calling</button>95 </div>96 </div>97 <div class="example-code" data-example></div>98 </section>99100 <section class="section wrap" aria-labelledby="features-title">101 <header class="section-head">102 <span class="eyebrow">Platform</span>103 <h2 id="features-title">Inference infrastructure, not a chat widget.</h2>104 <p>Everything below is implemented in the gateway today and described precisely in the documentation.</p>105 </header>106 <ul class="features">107 <li class="feature">108 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M8 7 3 12l5 5M16 7l5 5-5 5M13.500 4l-3 16"/></svg></span>109 <h3>OpenAI-compatible</h3>110 <p>Point the official OpenAI SDKs at a new base URL. Chat completions, text completions and model listing follow the same request and response shapes.</p>111 <a href="/docs/chat-completions" aria-label="OpenAI-compatible: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>112 </li>113 <li class="feature">114 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 7h10M3 12h18M3 17h6"/><circle cx="17" cy="7" r="1.200" fill="currentColor"/><circle cx="13" cy="17" r="1.200" fill="currentColor"/></svg></span>115 <h3>Token streaming</h3>116 <p>Server-sent events with keepalive comments during prefill, an optional usage chunk, and explicit in-stream error events.</p>117 <a href="/docs/streaming" aria-label="Token streaming: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>118 </li>119 <li class="feature">120 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><rect x="5" y="11" width="14" height="9" rx="2.500"/><path d="M8 11V8a4 4 0 0 1 8 0v3"/></svg></span>121 <h3>Private by construction</h3>122 <p>Inference runs on our own Apple Silicon hardware. No third-party model API is involved, and prompts and completions are never written to logs: only counters and timings are recorded.</p>123 <a href="/docs/overview#privacy" aria-label="Private by construction: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>124 </li>125 <li class="feature">126 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M14.500 6.500a4 4 0 0 0-5.300 5.300L4 17l3 3 5.200-5.200a4 4 0 0 0 5.300-5.300L15 12l-3-3 2.500-2.500Z"/></svg></span>127 <h3>Tool calling</h3>128 <p>Declare functions, receive structured tool calls, return results, continue the conversation. Each call arrives whole in a single chunk.</p>129 <a href="/docs/tool-calling" aria-label="Tool calling: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>130 </li>131 <li class="feature">132 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M8 4c-2 0-3 1-3 3v2c0 1.500-1 3-2 3 1 0 2 1.500 2 3v2c0 2 1 3 3 3M16 4c2 0 3 1 3 3v2c0 1.500 1 3 2 3-1 0-2 1.500-2 3v2c0 2-1 3-3 3"/><path d="m9.500 12 1.800 1.800 3.200-3.600"/></svg></span>133 <h3>Validated JSON</h3>134 <p>JSON object and JSON Schema modes are checked server-side, with one automatic repair attempt on non-streaming requests. Honest semantics: validated, not grammar-constrained.</p>135 <a href="/docs/structured-outputs" aria-label="Validated JSON: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>136 </li>137 <li class="feature">138 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><circle cx="8" cy="15" r="4"/><path d="m11 12 8-8M16 7l2.500 2.500M14 9l1.500 1.500"/></svg></span>139 <h3>Per-app keys and limits</h3>140 <p>Each application gets its own key, stored hashed, with scoped models and endpoints, request and token rates, daily and monthly quotas, and a concurrency cap.</p>141 <a href="/docs/authentication" aria-label="Per-app keys and limits: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>142 </li>143 <li class="feature">144 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M4 6h12M4 12h16M4 18h8"/><path d="m17 16 3 2-3 2"/></svg></span>145 <h3>Fair scheduling</h3>146 <p>Four priority classes with anti-starvation aging, a bounded queue, and backpressure that answers 503 with Retry-After instead of timing out silently.</p>147 <a href="/docs/rate-limits" aria-label="Fair scheduling: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>148 </li>149 <li class="feature">150 <span class="feature-icon"><svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><rect x="4" y="4" width="7" height="7" rx="1.500"/><rect x="13" y="4" width="7" height="7" rx="1.500"/><rect x="4" y="13" width="7" height="7" rx="1.500"/><path d="M13 16.500h7M16.500 13v7"/></svg></span>151 <h3>Batching and prefix cache</h3>152 <p>The MLX engine batches concurrent generations and reuses KV state for shared prompt prefixes. Cached prompt tokens are reported in every usage object.</p>153 <a href="/docs/performance" aria-label="Batching and prefix cache: read the documentation">Learn more <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>154 </li>155 </ul>156 </section>157158 <section class="section wrap" aria-labelledby="arch-title">159 <header class="section-head">160 <span class="eyebrow">Architecture</span>161 <h2 id="arch-title">A short, observable path from request to token.</h2>162 <p>Requests are authenticated, metered and queued by priority before they reach the MLX engine. The engine and the gateway run on the same Apple Silicon machine.</p>163 </header>164 <figure class="arch-figure" data-arch>165 <svg class="arch arch-wide" viewBox="0 0 1160 224" role="img" aria-labelledby="arch-t arch-d"><title id="arch-t">Request path</title><desc id="arch-d">Client apps connect over HTTPS to a TLS gateway, which forwards to the API Gateway (authentication, rate limits, priority queue). The API Gateway feeds the MLX engine (continuous batching, KV prefix cache), which runs on the Apple Silicon GPU and unified memory.</desc><defs>166<linearGradient id="aw-g" x1="0" y1="0" x2="1" y2="1"><stop offset="0" stop-color="#5ef2c4"/><stop offset=".5" stop-color="#4cc9f0"/><stop offset="1" stop-color="#9b8cff"/></linearGradient>167<marker id="aw-a" viewBox="0 0 10 10" refX="8" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path d="M1 1.5 8.5 5 1 8.5" fill="none" stroke="context-stroke" stroke-width="1.6" stroke-linecap="round" stroke-linejoin="round"/></marker>168</defs><g class="arch-node arch-client"><rect class="arch-box" x="0" y="30" width="192" height="164" rx="16"/><text class="arch-num" x="18" y="60">01</text><text class="arch-title" x="18" y="88">Client apps</text><circle class="arch-dot" cx="22" cy="112" r="2.5"/><text class="arch-line" x="34" y="116">OpenAI SDKs</text><circle class="arch-dot" cx="22" cy="134" r="2.5"/><text class="arch-line" x="34" y="138">cURL · URLSession</text><circle class="arch-dot" cx="22" cy="156" r="2.5"/><text class="arch-line" x="34" y="160">Your backend</text></g><path class="arch-link" d="M197 112.0H237" marker-end="url(#aw-a)"/><path class="arch-flow" d="M197 112.0H237"/><text class="arch-label" x="217.0" y="100.0" text-anchor="middle">HTTPS</text><g class="arch-node arch-tls"><rect class="arch-box" x="242" y="30" width="192" height="164" rx="16"/><text class="arch-num" x="260" y="60">02</text><text class="arch-title" x="260" y="88">HTTPS / TLS gateway</text><circle class="arch-dot" cx="264" cy="112" r="2.5"/><text class="arch-line" x="276" y="116">Public endpoint</text><circle class="arch-dot" cx="264" cy="134" r="2.5"/><text class="arch-line" x="276" y="138">TLS termination</text></g><path class="arch-link" d="M439 112.0H479" marker-end="url(#aw-a)"/><path class="arch-flow" d="M439 112.0H479"/><g class="arch-node arch-api"><rect class="arch-box" x="484" y="30" width="192" height="164" rx="16"/><text class="arch-num" x="502" y="60">03</text><text class="arch-title" x="502" y="88">API Gateway</text><circle class="arch-dot" cx="506" cy="112" r="2.5"/><text class="arch-line" x="518" y="116">Authentication</text><circle class="arch-dot" cx="506" cy="134" r="2.5"/><text class="arch-line" x="518" y="138">Rate limits</text><circle class="arch-dot" cx="506" cy="156" r="2.5"/><text class="arch-line" x="518" y="160">Priority queue</text></g><path class="arch-link" d="M681 112.0H721" marker-end="url(#aw-a)"/><path class="arch-flow" d="M681 112.0H721"/><g class="arch-node arch-mlx"><rect class="arch-box" x="726" y="30" width="192" height="164" rx="16"/><text class="arch-num" x="744" y="60">04</text><text class="arch-title" x="744" y="88">MLX engine</text><circle class="arch-dot" cx="748" cy="112" r="2.5"/><text class="arch-line" x="760" y="116">Continuous batching</text><circle class="arch-dot" cx="748" cy="134" r="2.5"/><text class="arch-line" x="760" y="138">KV prefix cache</text></g><path class="arch-link" d="M923 112.0H963" marker-end="url(#aw-a)"/><path class="arch-flow" d="M923 112.0H963"/><text class="arch-label" x="943.0" y="100.0" text-anchor="middle">Metal</text><g class="arch-node arch-gpu"><rect class="arch-box" x="968" y="30" width="192" height="164" rx="16"/><text class="arch-num" x="986" y="60">05</text><text class="arch-title" x="986" y="88">Apple Silicon</text><circle class="arch-dot" cx="990" cy="112" r="2.5"/><text class="arch-line" x="1002" y="116">GPU (Metal)</text><circle class="arch-dot" cx="990" cy="134" r="2.5"/><text class="arch-line" x="1002" y="138">Unified memory</text></g><text class="arch-caption" x="484" y="216">gateway process</text><text class="arch-caption" x="726" y="216">engine process · same machine</text></svg>169 <svg class="arch arch-tall" viewBox="0 0 320 834" role="img" aria-labelledby="archv-t archv-d"><title id="archv-t">Request path</title><desc id="archv-d">Client apps connect over HTTPS to a TLS gateway, which forwards to the API Gateway (authentication, rate limits, priority queue). The API Gateway feeds the MLX engine (continuous batching, KV prefix cache), which runs on the Apple Silicon GPU and unified memory.</desc><defs>170<linearGradient id="av-g" x1="0" y1="0" x2="1" y2="1"><stop offset="0" stop-color="#5ef2c4"/><stop offset=".5" stop-color="#4cc9f0"/><stop offset="1" stop-color="#9b8cff"/></linearGradient>171<marker id="av-a" viewBox="0 0 10 10" refX="8" refY="5" markerWidth="7" markerHeight="7" orient="auto-start-reverse"><path d="M1 1.5 8.5 5 1 8.5" fill="none" stroke="context-stroke" stroke-width="1.6" stroke-linecap="round" stroke-linejoin="round"/></marker>172</defs><g class="arch-node arch-client"><rect class="arch-box" x="1" y="2" width="318" height="152" rx="16"/><text class="arch-num" x="19" y="32">01</text><text class="arch-title" x="19" y="60">Client apps</text><circle class="arch-dot" cx="23" cy="84" r="2.5"/><text class="arch-line" x="35" y="88">OpenAI SDKs</text><circle class="arch-dot" cx="23" cy="106" r="2.5"/><text class="arch-line" x="35" y="110">cURL · URLSession</text><circle class="arch-dot" cx="23" cy="128" r="2.5"/><text class="arch-line" x="35" y="132">Your backend</text></g><path class="arch-link" d="M160.0 159V183" marker-end="url(#av-a)"/><path class="arch-flow" d="M160.0 159V183"/><text class="arch-label" x="172.0" y="175.0">HTTPS</text><g class="arch-node arch-tls"><rect class="arch-box" x="1" y="188" width="318" height="130" rx="16"/><text class="arch-num" x="19" y="218">02</text><text class="arch-title" x="19" y="246">HTTPS / TLS gateway</text><circle class="arch-dot" cx="23" cy="270" r="2.5"/><text class="arch-line" x="35" y="274">Public endpoint</text><circle class="arch-dot" cx="23" cy="292" r="2.5"/><text class="arch-line" x="35" y="296">TLS termination</text></g><path class="arch-link" d="M160.0 323V347" marker-end="url(#av-a)"/><path class="arch-flow" d="M160.0 323V347"/><g class="arch-node arch-api"><rect class="arch-box" x="1" y="352" width="318" height="152" rx="16"/><text class="arch-num" x="19" y="382">03</text><text class="arch-title" x="19" y="410">API Gateway</text><circle class="arch-dot" cx="23" cy="434" r="2.5"/><text class="arch-line" x="35" y="438">Authentication</text><circle class="arch-dot" cx="23" cy="456" r="2.5"/><text class="arch-line" x="35" y="460">Rate limits</text><circle class="arch-dot" cx="23" cy="478" r="2.5"/><text class="arch-line" x="35" y="482">Priority queue</text></g><path class="arch-link" d="M160.0 509V533" marker-end="url(#av-a)"/><path class="arch-flow" d="M160.0 509V533"/><g class="arch-node arch-mlx"><rect class="arch-box" x="1" y="538" width="318" height="130" rx="16"/><text class="arch-num" x="19" y="568">04</text><text class="arch-title" x="19" y="596">MLX engine</text><circle class="arch-dot" cx="23" cy="620" r="2.5"/><text class="arch-line" x="35" y="624">Continuous batching</text><circle class="arch-dot" cx="23" cy="642" r="2.5"/><text class="arch-line" x="35" y="646">KV prefix cache</text></g><path class="arch-link" d="M160.0 673V697" marker-end="url(#av-a)"/><path class="arch-flow" d="M160.0 673V697"/><text class="arch-label" x="172.0" y="689.0">Metal</text><g class="arch-node arch-gpu"><rect class="arch-box" x="1" y="702" width="318" height="130" rx="16"/><text class="arch-num" x="19" y="732">05</text><text class="arch-title" x="19" y="760">Apple Silicon</text><circle class="arch-dot" cx="23" cy="784" r="2.5"/><text class="arch-line" x="35" y="788">GPU (Metal)</text><circle class="arch-dot" cx="23" cy="806" r="2.5"/><text class="arch-line" x="35" y="810">Unified memory</text></g></svg>173 </figure>174 </section>175176 <section class="section wrap" aria-labelledby="cta-title">177 <div class="cta">178 <div>179 <h2 id="cta-title">Start with one request.</h2>180 <p>Read the five-minute guide, or try the model against the real API in the Playground.</p>181 </div>182 <div class="cta-actions">183 <a class="btn btn-primary" href="/docs/getting-started">Read the docs <svg class="arrow" viewBox="0 0 16 16" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M3 8h10M9 4l4 4-4 4"/></svg></a>184 <a class="btn" href="/playground">Playground</a>185 </div>186 </div>187 </section>188 </main>189 <footer class="site-footer">190 <div class="wrap">191 <div class="footer-brand">192 <a class="brand" href="/" aria-label="MacLustr AI — home"><img src="/assets/logo.svg" alt="" width="28" height="28"><span>MacLustr <span>AI</span></span></a>193 <p>Local LLM inference on Apple Silicon with MLX, behind an OpenAI-compatible API.</p>194 </div>195 <nav aria-label="Documentation links">196 <h2>Build</h2>197 <ul>198 <li><a href="/docs/getting-started">Getting started</a></li>199 <li><a href="/docs/chat-completions">Chat completions</a></li>200 <li><a href="/docs/streaming">Streaming</a></li>201 <li><a href="/docs/sdk-examples">SDK examples</a></li>202 </ul>203 </nav>204 <nav aria-label="Reference links">205 <h2>Reference</h2>206 <ul>207 <li><a href="/docs/authentication">Authentication</a></li>208 <li><a href="/docs/error-handling">Error handling</a></li>209 <li><a href="/docs/rate-limits">Rate limits</a></li>210 <li><a href="/openapi.json">OpenAPI contract</a></li>211 </ul>212 </nav>213 <nav aria-label="Service links">214 <h2>Service</h2>215 <ul>216 <li><a href="/playground">Playground</a></li>217 <li><a href="/status">Status</a></li>218 <li><a href="/changelog">Changelog</a></li>219 <li><a href="/docs/performance">Performance</a></li>220 </ul>221 </nav>222 <div class="footer-meta">223 <span>© <span data-year>2026</span> MacLustr. Inference runs on our own hardware.</span>224 <span>Base URL <code>https://llm.spboucher.ai/v1</code></span>225 </div>226 </div>227 </footer>228 <script type="module" src="/assets/site/landing.js"></script>229</body>230</html>231