From c15ec320c919686fcf070b202957a24f58147c34 Mon Sep 17 00:00:00 2001
From: Andrea Griffiths <andreagriffiths11@users.noreply.github.com>
Date: Thu, 11 Jun 2026 08:16:51 -0700
Subject: [PATCH] docs(examples): add github copilot byok section (cli + vs
 code)

Adds a new section under /docs/examples covering Copilot BYOK:
- copilot cli: env vars (COPILOT_PROVIDER_BASE_URL, COPILOT_PROVIDER_API_KEY,
  COPILOT_MODEL) targeting api.nan.builders/v1.
- vs code: step-by-step using the Custom Endpoint provider plus a ready-to-paste
  chatLanguageModels.json snippet.

Notes models that satisfy Copilot's tool-calling + streaming requirement
(qwen3.6, deepseek-v4-flash, mimo-v2.5).

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 src/pages/docs/examples.astro | 48 +++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/src/pages/docs/examples.astro b/src/pages/docs/examples.astro
index 0a580d9..b9726ea 100644
--- a/src/pages/docs/examples.astro
+++ b/src/pages/docs/examples.astro
@@ -234,6 +234,54 @@ for await (const chunk of stream) {
     Config para <code>~/.config/zed/settings.json</code> — incluye inline predictions.
   </p>
 
+  <!-- github copilot -->
+  <h2 id="copilot">github copilot (byok)</h2>
+  <p class="font-mono text-xs text-neutral-500 mb-6">cli + vs code · trae tu propia key</p>
+
+  <p>Copilot soporta endpoints OpenAI-compatibles via BYOK (bring your own key).
+    Usa un modelo con tool calling + streaming: <code>qwen3.6</code>,
+    <code>deepseek-v4-flash</code> o <code>mimo-v2.5</code>.</p>
+
+  <h3 id="copilot-cli">copilot cli</h3>
+  <CodeBlock language="bash" code={`export COPILOT_PROVIDER_BASE_URL=https://api.nan.builders/v1
+export COPILOT_PROVIDER_API_KEY=sk-tu-key-aqui
+export COPILOT_MODEL=qwen3.6
+copilot`} />
+  <p class="font-mono text-xs text-neutral-500 mb-5">
+    Variables de entorno antes de lanzar <code>copilot</code>. Override por sesión con <code>--model qwen3.6</code>. Docs: <a href="https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/use-byok-models" target="_blank" rel="noopener noreferrer" class="text-violet-400 hover:text-violet-300 underline underline-offset-2">use-byok-models</a>.
+  </p>
+
+  <h3 id="copilot-vscode">vs code</h3>
+  <ol class="list-decimal list-inside text-sm text-neutral-300 leading-relaxed mb-5 space-y-1">
+    <li>Abrir la paleta de comandos (<code>⌘⇧P</code> / <code>Ctrl+Shift+P</code>) → <code>Chat: Manage Language Models</code>.</li>
+    <li><code>Add Models</code> → <code>Custom Endpoint</code>.</li>
+    <li>Group name: <code>NaN</code>. Display name: <code>qwen3.6</code>. API key: tu <code>sk-...</code>. API type: <code>Chat Completions</code>.</li>
+    <li>VS Code abre <code>chatLanguageModels.json</code>. Reemplazar el contenido con el JSON de abajo y guardar.</li>
+    <li>Seleccionar <code>qwen3.6</code> en el model picker del chat.</li>
+  </ol>
+  <CodeBlock language="json" code={`[
+  {
+    "name": "NaN",
+    "vendor": "customendpoint",
+    "apiKey": "sk-tu-key-aqui",
+    "apiType": "chat-completions",
+    "models": [
+      {
+        "id": "qwen3.6",
+        "name": "qwen3.6",
+        "url": "https://api.nan.builders/v1/chat/completions",
+        "toolCalling": true,
+        "vision": true,
+        "maxInputTokens": 262144,
+        "maxOutputTokens": 16384
+      }
+    ]
+  }
+]`} />
+  <p class="font-mono text-xs text-neutral-500 mb-5">
+    Para agent mode usa un modelo con tool calling (<code>qwen3.6</code>, <code>deepseek-v4-flash</code>, <code>mimo-v2.5</code>). Si el modelo no aparece en el picker, reiniciar VS Code. Docs: <a href="https://code.visualstudio.com/docs/copilot/customization/language-models#_bring-your-own-language-model-key" target="_blank" rel="noopener noreferrer" class="text-violet-400 hover:text-violet-300 underline underline-offset-2">language-models</a>.
+  </p>
+
   <!-- qwen3-embedding -->
   <h2 id="qwen3-embedding">modelo: qwen3-embedding</h2>
   <p class="font-mono text-xs text-neutral-500 mb-6">embeddings vectoriales</p>