doccupine 0.0.63 → 0.0.65

package/dist/templates/components/layout/Update.js

@@ -46,7 +46,7 @@ const StyledUpdateChildren = styled.div\`
   gap: 20px;
 \`;
 
-export interface UpdateProps extends React.HTMLAttributes<HTMLDivElement> {
+interface UpdateProps extends React.HTMLAttributes<HTMLDivElement> {
   children: React.ReactNode;
   label: string;
   description: string;

package/dist/templates/mdx/ai-assistant.mdx.d.ts

@@ -1 +1 @@
-export declare const aiAssistantMdxTemplate = "---\ntitle: \"AI Assistant\"\ndescription: \"Integrate AI capabilities into your Doccupine documentation using OpenAI, Anthropic, or Google Gemini.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 8\n---\n# AI Assistant\nDoccupine supports AI integration to enhance your documentation experience. You can use OpenAI, Anthropic, or Google Gemini to power AI features in your documentation site. The AI assistant uses your documentation content as context, allowing users to ask questions about your docs and receive accurate answers based on the documentation.\n\n## Setup\nTo enable AI features, create an `.env` file in the directory where your website is generated. By default, this is the `nextjs-app/` directory.\n\n## Configuration\nCreate an `.env` file with the following configuration options:\n\n```env\n# LLM Provider Configuration\n# Choose your preferred LLM provider: openai, anthropic, or google\nLLM_PROVIDER=openai\n\n# API Keys (set the one matching your provider)\nOPENAI_API_KEY=your_openai_api_key_here\nANTHROPIC_API_KEY=your_anthropic_api_key_here\nGOOGLE_API_KEY=your_google_api_key_here\n\n# Optional: Override default chat model (see your provider's docs for available models)\n# LLM_CHAT_MODEL=your-model-id\n\n# Optional: Override default embedding model (see your provider's docs for available models)\n# Note: Anthropic doesn't provide embeddings, will fallback to OpenAI\n# LLM_EMBEDDING_MODEL=your-embedding-model-id\n\n# Optional: Set temperature (0-1, default: 0)\n# LLM_TEMPERATURE=0\n```\n\n## Provider Selection\nSet `LLM_PROVIDER` to one of the following values:\n- `openai` - Use OpenAI's models\n- `anthropic` - Use Anthropic's models\n- `google` - Use Google's models\n\n## API Keys\nYou need to set the API key that matches your chosen provider:\n- For OpenAI: Set `OPENAI_API_KEY`\n- For Anthropic: Set `ANTHROPIC_API_KEY`\n- For Google: Set `GOOGLE_API_KEY`\n\n<Callout type=\"warning\">\n  Keep your API keys secure. Never commit your `.env` file to version control.\n</Callout>\n\n<Callout type=\"note\">\n  Doccupine automatically adds `.env` to your `.gitignore` file.\n</Callout>\n\n## Using Anthropic with OpenAI\nIf you want to use Anthropic as your LLM provider, you must also have an OpenAI API key set. Here's why:\n\n### The Situation\nAnthropic (Claude) does not provide an embeddings API. They only offer chat/completion models, not text embeddings.\n\nYour RAG (Retrieval-Augmented Generation) system has two components:\n- **Chat/Completion** - Generates answers, works with Anthropic.\n- **Embeddings** - Creates vector representations of text for search, Anthropic doesn't provide this.\n\nWhen using Anthropic as your `LLM_PROVIDER`, Doccupine will use Anthropic for chat/completion tasks, but will automatically fallback to OpenAI for embeddings. This means you need both API keys configured:\n\n```env\nLLM_PROVIDER=anthropic\nANTHROPIC_API_KEY=your_anthropic_api_key_here\nOPENAI_API_KEY=your_openai_api_key_here\n```\n\nThis hybrid approach allows you to leverage Anthropic's powerful chat models while still having access to embeddings functionality through OpenAI.\n\n## Optional Settings\n\n### Chat Model\nOverride the default chat model by uncommenting and setting `LLM_CHAT_MODEL`. You can use any available model from your chosen provider. For a complete list of available models, refer to the official documentation:\n- [OpenAI Models](https://platform.openai.com/docs/models)\n- [Anthropic Models](https://docs.anthropic.com/claude/docs/models-overview)\n- [Google Gemini Models](https://ai.google.dev/models/gemini)\n\n### Embedding Model\nOverride the default embedding model by uncommenting and setting `LLM_EMBEDDING_MODEL`. For a complete list of available embedding models, refer to the official documentation:\n- [OpenAI Embeddings](https://platform.openai.com/docs/guides/embeddings)\n- [Google Gemini Embeddings](https://ai.google.dev/gemini-api/docs/embeddings)\n- **Anthropic**: Anthropic doesn't provide embeddings. If you use Anthropic as your provider, Doccupine will fallback to OpenAI for embeddings.\n\n### Temperature\nControl the randomness of AI responses by setting `LLM_TEMPERATURE` to a value between 0 and 1:\n- `0` - More deterministic and focused responses (default)\n- `1` - More creative and varied responses";
+export declare const aiAssistantMdxTemplate = "---\ntitle: \"AI Assistant\"\ndescription: \"Integrate AI capabilities into your Doccupine documentation using OpenAI, Anthropic, or Google Gemini.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 8\n---\n# AI Assistant\nDoccupine supports AI integration to enhance your documentation experience. You can use OpenAI, Anthropic, or Google Gemini to power AI features in your documentation site. The AI assistant uses your documentation content as context, allowing users to ask questions about your docs and receive accurate answers based on the documentation.\n\n## Setup\nTo enable AI features, create an `.env` file in the directory where your website is generated. By default, this is the `nextjs-app/` directory.\n\n## Configuration\nCreate an `.env` file with the following configuration options:\n\n```env\n# LLM Provider Configuration\n# Choose your preferred LLM provider: openai, anthropic, or google\nLLM_PROVIDER=openai\n\n# API Keys (set the one matching your provider)\nOPENAI_API_KEY=your_openai_api_key_here\nANTHROPIC_API_KEY=your_anthropic_api_key_here\nGOOGLE_API_KEY=your_google_api_key_here\n\n# Optional: Override default chat model (see your provider's docs for available models)\n# LLM_CHAT_MODEL=your-model-id\n\n# Optional: Override default embedding model (see your provider's docs for available models)\n# Note: Anthropic doesn't provide embeddings, will fallback to OpenAI\n# LLM_EMBEDDING_MODEL=your-embedding-model-id\n\n# Optional: Set temperature (0-1, default: 0)\n# LLM_TEMPERATURE=0\n```\n\n## Provider Selection\nSet `LLM_PROVIDER` to one of the following values:\n- `openai` - Use OpenAI's models\n- `anthropic` - Use Anthropic's models\n- `google` - Use Google's models\n\n## API Keys\nYou need to set the API key that matches your chosen provider:\n- For OpenAI: Set `OPENAI_API_KEY`\n- For Anthropic: Set `ANTHROPIC_API_KEY`\n- For Google: Set `GOOGLE_API_KEY`\n\n<Callout type=\"warning\">\n  Keep your API keys secure. Never commit your `.env` file to version control.\n</Callout>\n\n<Callout type=\"note\">\n  Doccupine automatically adds `.env` to your `.gitignore` file.\n</Callout>\n\n## Using Anthropic with OpenAI\nIf you want to use Anthropic as your LLM provider, you must also have an OpenAI API key set. Here's why:\n\n### The Situation\nAnthropic (Claude) does not provide an embeddings API. They only offer chat/completion models, not text embeddings.\n\nYour RAG (Retrieval-Augmented Generation) system has two components:\n- **Chat/Completion** - Generates answers, works with Anthropic.\n- **Embeddings** - Creates vector representations of text for search, Anthropic doesn't provide this.\n\nWhen using Anthropic as your `LLM_PROVIDER`, Doccupine will use Anthropic for chat/completion tasks, but will automatically fallback to OpenAI for embeddings. This means you need both API keys configured:\n\n```env\nLLM_PROVIDER=anthropic\nANTHROPIC_API_KEY=your_anthropic_api_key_here\nOPENAI_API_KEY=your_openai_api_key_here\n```\n\nThis hybrid approach allows you to leverage Anthropic's powerful chat models while still having access to embeddings functionality through OpenAI.\n\n## Default models\n\n| Provider | Chat model | Embedding model |\n|---|---|---|\n| OpenAI | `gpt-4.1-nano` | `text-embedding-3-small` |\n| Anthropic | `claude-sonnet-4-5-20250929` | OpenAI fallback |\n| Google | `gemini-2.5-flash-lite` | `gemini-embedding-001` |\n\n## Optional Settings\n\n### Chat Model\nOverride the default chat model by uncommenting and setting `LLM_CHAT_MODEL`. You can use any available model from your chosen provider. For a complete list of available models, refer to the official documentation:\n- [OpenAI Models](https://platform.openai.com/docs/models)\n- [Anthropic Models](https://docs.anthropic.com/claude/docs/models-overview)\n- [Google Gemini Models](https://ai.google.dev/models/gemini)\n\n### Embedding Model\nOverride the default embedding model by uncommenting and setting `LLM_EMBEDDING_MODEL`. For a complete list of available embedding models, refer to the official documentation:\n- [OpenAI Embeddings](https://platform.openai.com/docs/guides/embeddings)\n- [Google Gemini Embeddings](https://ai.google.dev/gemini-api/docs/embeddings)\n- **Anthropic**: Anthropic doesn't provide embeddings. If you use Anthropic as your provider, Doccupine will fallback to OpenAI for embeddings.\n\n### Temperature\nControl the randomness of AI responses by setting `LLM_TEMPERATURE` to a value between 0 and 1:\n- `0` - More deterministic and focused responses (default)\n- `1` - More creative and varied responses";

package/dist/templates/mdx/ai-assistant.mdx.js

@@ -76,6 +76,14 @@ OPENAI_API_KEY=your_openai_api_key_here
 
 This hybrid approach allows you to leverage Anthropic's powerful chat models while still having access to embeddings functionality through OpenAI.
 
+## Default models
+
+| Provider | Chat model | Embedding model |
+|---|---|---|
+| OpenAI | \`gpt-4.1-nano\` | \`text-embedding-3-small\` |
+| Anthropic | \`claude-sonnet-4-5-20250929\` | OpenAI fallback |
+| Google | \`gemini-2.5-flash-lite\` | \`gemini-embedding-001\` |
+
 ## Optional Settings
 
 ### Chat Model

package/dist/templates/mdx/callouts.mdx.d.ts

@@ -1 +1 @@
-export declare const calloutsMdxTemplate = "---\ntitle: \"Callouts\"\ndescription: \"Make your content stand out by using callouts for extra emphasis.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 7\n---\n# Callouts\nMake your content stand out by using callouts for extra emphasis.\n\nYou can format them as Note, Warning, Info, Danger and Success.\n\n## Callouts Usage\nYou can use the Callouts component directly within your MDX files without any import. The following example shows a basic usage:\n\n~~~mdx\n<Callout type=\"note\">\n  This is a note callout\n</Callout>\n\n<Callout type=\"warning\">\n  This is a warning callout\n</Callout>\n\n<Callout type=\"info\">\n  This is an info callout\n</Callout>\n\n<Callout type=\"danger\">\n  This is a danger callout\n</Callout>\n\n<Callout type=\"success\">\n This is a success callout\n</Callout>\n~~~\n\n<Callout type=\"note\">\n  This is a note callout\n</Callout>\n\n<Callout type=\"warning\">\n  This is a warning callout\n</Callout>\n\n<Callout type=\"info\">\n  This is an info callout\n</Callout>\n\n<Callout type=\"danger\">\n  This is a danger callout\n</Callout>\n\n<Callout type=\"success\">\n This is a success callout\n</Callout>\n\n## Properties\n\n<Field value=\"type\" type=\"string\" required>\n  The type of the callout.\n</Field>\n\n<Field value=\"children\" type=\"node\" required>\n  The content of the callout.\n</Field>";
+export declare const calloutsMdxTemplate = "---\ntitle: \"Callouts\"\ndescription: \"Make your content stand out by using callouts for extra emphasis.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 7\n---\n# Callouts\nMake your content stand out by using callouts for extra emphasis.\n\nYou can format them as Note, Warning, Info, Danger and Success.\n\n## Callouts Usage\nYou can use the Callouts component directly within your MDX files without any import. The following example shows a basic usage:\n\n~~~mdx\n<Callout type=\"note\">\n  This is a note callout\n</Callout>\n\n<Callout type=\"warning\">\n  This is a warning callout\n</Callout>\n\n<Callout type=\"info\">\n  This is an info callout\n</Callout>\n\n<Callout type=\"danger\">\n  This is a danger callout\n</Callout>\n\n<Callout type=\"success\">\n This is a success callout\n</Callout>\n~~~\n\n<Callout type=\"note\">\n  This is a note callout\n</Callout>\n\n<Callout type=\"warning\">\n  This is a warning callout\n</Callout>\n\n<Callout type=\"info\">\n  This is an info callout\n</Callout>\n\n<Callout type=\"danger\">\n  This is a danger callout\n</Callout>\n\n<Callout type=\"success\">\n This is a success callout\n</Callout>\n\n## Properties\n\n<Field value=\"type\" type=\"string\">\n  The type of the callout: `note`, `info`, `warning`, `danger`, or `success`.\n</Field>\n\n<Field value=\"icon\" type=\"string\">\n  A custom [Lucide](https://lucide.dev/icons) icon name. Overrides the default icon for the callout type.\n</Field>\n\n<Field value=\"children\" type=\"node\" required>\n  The content of the callout.\n</Field>";

package/dist/templates/mdx/callouts.mdx.js

@@ -58,8 +58,12 @@ You can use the Callouts component directly within your MDX files without any im
 
 ## Properties
 
-<Field value="type" type="string" required>
-  The type of the callout.
+<Field value="type" type="string">
+  The type of the callout: \`note\`, \`info\`, \`warning\`, \`danger\`, or \`success\`.
+</Field>
+
+<Field value="icon" type="string">
+  A custom [Lucide](https://lucide.dev/icons) icon name. Overrides the default icon for the callout type.
 </Field>
 
 <Field value="children" type="node" required>

package/dist/templates/mdx/cards.mdx.d.ts

@@ -1 +1 @@
-export declare const cardsMdxTemplate = "---\ntitle: \"Cards\"\ndescription: \"Cards act as visual containers for your content, giving you flexibility to combine text, icons, images, and links in a clean and organized way.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 6\n---\n# Cards\nDuplicate a page or section with ease, then emphasize important information or links using customizable layouts and icons.\n\nCards act as visual containers for your content, giving you flexibility to combine text, icons, images, and links in a clean and organized way.\n\n## Cards Usage\nYou can use the Cards component directly within your MDX files without any import. The following example shows a basic usage:\n\n~~~mdx\n<Card title=\"Note\" icon=\"badge-info\">\n  Doccupine CLI is a command-line tool that helps you create and manage your Doccupine project. It provides a simple and intuitive interface for creating and configuring your project.\n</Card>\n~~~\n\n<Card title=\"Note\" icon=\"badge-info\">\n  Doccupine CLI is a command-line tool that helps you create and manage your Doccupine project. It provides a simple and intuitive interface for creating and configuring your project.\n</Card>\n\n## Properties\n\n<Field value=\"title\" type=\"string\" required>\n  The title of the card.\n</Field>\n\n<Field value=\"icon\" type=\"string\" required>\n  The icon to display in the card.\n</Field>\n\n- [**Lucide icon**](https://lucide.dev/icons) name\n\n<Field value=\"children\" type=\"node\" required>\n  The content of the card.\n</Field>";
+export declare const cardsMdxTemplate = "---\ntitle: \"Cards\"\ndescription: \"Cards act as visual containers for your content, giving you flexibility to combine text, icons, images, and links in a clean and organized way.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 6\n---\n# Cards\nDuplicate a page or section with ease, then emphasize important information or links using customizable layouts and icons.\n\nCards act as visual containers for your content, giving you flexibility to combine text, icons, images, and links in a clean and organized way.\n\n## Cards Usage\nYou can use the Cards component directly within your MDX files without any import. The following example shows a basic usage:\n\n~~~mdx\n<Card title=\"Note\" icon=\"badge-info\">\n  Doccupine CLI is a command-line tool that helps you create and manage your Doccupine project. It provides a simple and intuitive interface for creating and configuring your project.\n</Card>\n~~~\n\n<Card title=\"Note\" icon=\"badge-info\">\n  Doccupine CLI is a command-line tool that helps you create and manage your Doccupine project. It provides a simple and intuitive interface for creating and configuring your project.\n</Card>\n\n## Link Card\n\nPass a `href` prop to turn the card into a clickable link. The card will display interactive hover and focus styles automatically.\n\n~~~mdx\n<Card title=\"Getting Started\" icon=\"rocket\" href=\"/cards\">\n  Learn how to set up Doccupine and create your first documentation site.\n</Card>\n~~~\n\n<Card title=\"Getting Started\" icon=\"rocket\" href=\"/cards\">\n  Learn how to set up Doccupine and create your first documentation site.\n</Card>\n\n## Properties\n\n<Field value=\"title\" type=\"string\" required>\n  The title of the card.\n</Field>\n\n<Field value=\"icon\" type=\"string\">\n  The [Lucide](https://lucide.dev/icons) icon name to display in the card.\n</Field>\n\n<Field value=\"href\" type=\"string\">\n  A URL or path to link to. When provided, the card becomes a clickable link with interactive styles.\n</Field>\n\n<Field value=\"children\" type=\"node\" required>\n  The content of the card.\n</Field>";

package/dist/templates/mdx/cards.mdx.js

@@ -24,17 +24,33 @@ You can use the Cards component directly within your MDX files without any impor
   Doccupine CLI is a command-line tool that helps you create and manage your Doccupine project. It provides a simple and intuitive interface for creating and configuring your project.
 </Card>
 
+## Link Card
+
+Pass a \`href\` prop to turn the card into a clickable link. The card will display interactive hover and focus styles automatically.
+
+~~~mdx
+<Card title="Getting Started" icon="rocket" href="/cards">
+  Learn how to set up Doccupine and create your first documentation site.
+</Card>
+~~~
+
+<Card title="Getting Started" icon="rocket" href="/cards">
+  Learn how to set up Doccupine and create your first documentation site.
+</Card>
+
 ## Properties
 
 <Field value="title" type="string" required>
   The title of the card.
 </Field>
 
-<Field value="icon" type="string" required>
-  The icon to display in the card.
+<Field value="icon" type="string">
+  The [Lucide](https://lucide.dev/icons) icon name to display in the card.
 </Field>
 
-- [**Lucide icon**](https://lucide.dev/icons) name
+<Field value="href" type="string">
+  A URL or path to link to. When provided, the card becomes a clickable link with interactive styles.
+</Field>
 
 <Field value="children" type="node" required>
   The content of the card.

package/dist/templates/mdx/columns.mdx.d.ts

@@ -1 +1 @@
-export declare const columnsMdxTemplate = "---\ntitle: \"Columns\"\ndescription: \"Columns are used to organize content in a grid-like structure.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 12\n---\n# Columns\nArrange multiple cards neatly in a side-by-side grid layout.\n\nThe `Columns` component helps you organize several `Card` elements into a visually balanced grid. By choosing how many columns you want, you can control the layout and spacing of your cards.\n\n## Columns Usage\nYou can use the `Columns` component to create a grid of cards with a specified number of columns.\n\n```mdx\n<Columns cols={2}>\n  <Card title=\"Getting Started\" icon=\"rocket\">\n    Kick off your project using our easy quickstart guide.\n  </Card>\n\n  <Card title=\"API Reference\" icon=\"code\">\n    Browse all endpoints, parameters, and code examples for your API integration.\n  </Card>\n</Columns>\n```\n\n<Columns cols={2}>\n  <Card title=\"Getting Started\" icon=\"rocket\">\n    Kick off your project using our easy quickstart guide.\n  </Card>\n\n  <Card title=\"API Reference\" icon=\"code\">\n    Browse all endpoints, parameters, and code examples for your API integration.\n  </Card>\n</Columns>\n\n## Properties\n\n<Field value=\"cols\" type=\"number\" required>\n  The number of columns in the grid.\n</Field>\n\n<Field value=\"children\" type=\"node\" required>\n  The content of the columns.\n</Field>";
+export declare const columnsMdxTemplate = "---\ntitle: \"Columns\"\ndescription: \"Columns are used to organize content in a grid-like structure.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 12\n---\n# Columns\nArrange multiple cards neatly in a side-by-side grid layout.\n\nThe `Columns` component helps you organize several `Card` elements into a visually balanced grid. By choosing how many columns you want, you can control the layout and spacing of your cards.\n\n## Columns Usage\nYou can use the `Columns` component to create a grid of cards with a specified number of columns.\n\n```mdx\n<Columns cols={2}>\n  <Card title=\"Getting Started\" icon=\"rocket\">\n    Kick off your project using our easy quickstart guide.\n  </Card>\n\n  <Card title=\"API Reference\" icon=\"code\">\n    Browse all endpoints, parameters, and code examples for your API integration.\n  </Card>\n</Columns>\n```\n\n<Columns cols={2}>\n  <Card title=\"Getting Started\" icon=\"rocket\">\n    Kick off your project using our easy quickstart guide.\n  </Card>\n\n  <Card title=\"API Reference\" icon=\"code\">\n    Browse all endpoints, parameters, and code examples for your API integration.\n  </Card>\n</Columns>\n\n## Properties\n\n<Field value=\"cols\" type=\"number\">\n  The number of columns in the grid. Defaults to 1.\n</Field>\n\n<Field value=\"children\" type=\"node\" required>\n  The content of the columns.\n</Field>";

package/dist/templates/mdx/columns.mdx.js

@@ -38,8 +38,8 @@ You can use the \`Columns\` component to create a grid of cards with a specified
 
 ## Properties
 
-<Field value="cols" type="number" required>
-  The number of columns in the grid.
+<Field value="cols" type="number">
+  The number of columns in the grid. Defaults to 1.
 </Field>
 
 <Field value="children" type="node" required>

package/dist/templates/mdx/commands.mdx.d.ts

@@ -1 +1 @@
-export declare const commandsMdxTemplate = "---\ntitle: \"Commands\"\ndescription: \"In this page, you can find all the commands available in Doccupine CLI.\"\ndate: \"2026-02-19\"\ncategory: \"Getting Started\"\ncategoryOrder: 0\norder: 1\n---\n# Commands\nIn this page, you can find all the commands available in Doccupine CLI.\n\n## Run Doccupine CLI\nCreate a new directory for your project and navigate to it in your terminal. Run the following command to create a new Doccupine project:\n\n```bash\nnpx doccupine\n```\n\nOnce you run the command, Doccupine will ask you to select a directory to store your MDX files. Choose the directory where you want to create your documentation files.\nAfter selecting the directory, Doccupine will ask you to enter the name of the directory for the generated website. Enter the name of the directory where you want to create your website.\n\nThis will start the development server on port 3000. Open your browser and navigate to http://localhost:3000 to view your documentation.\n\n## Verbose mode\n\n```bash\nnpx doccupine --verbose\n```\n\n<Callout type=\"note\">\n  This will show Next.js errors in the terminal console logs, providing more detailed output useful for debugging during development.\n</Callout>\n\n## Generate the website\n\n```bash\nnpx doccupine build\n```\n\nThis will generate the build files for your documentation website without starting the development server. You can then deploy the generated files to a hosting service of your choice.\n\n## Show current configuration\n\n```bash\nnpx doccupine config --show\n```\n\nThis will show the current configuration for Doccupine.\n\n## Reset configuration\n\n```bash\nnpx doccupine config --reset\n```\n\nThis will reset the current configuration for Doccupine.";
+export declare const commandsMdxTemplate = "---\ntitle: \"Commands\"\ndescription: \"In this page, you can find all the commands available in Doccupine CLI.\"\ndate: \"2026-02-19\"\ncategory: \"Getting Started\"\ncategoryOrder: 0\norder: 1\n---\n# Commands\nIn this page, you can find all the commands available in Doccupine CLI.\n\n## Run Doccupine CLI\nCreate a new directory for your project and navigate to it in your terminal. Run the following command to create a new Doccupine project:\n\n```bash\nnpx doccupine\n```\n\nOnce you run the command, Doccupine will ask you to select a directory to store your MDX files. Choose the directory where you want to create your documentation files.\nAfter selecting the directory, Doccupine will ask you to enter the name of the directory for the generated website. Enter the name of the directory where you want to create your website.\n\nThis will start the development server on port 3000. Open your browser and navigate to http://localhost:3000 to view your documentation.\n\n## Options\n\n| Flag | Description |\n|---|---|\n| `--port <port>` | Port for the dev server (default: `3000`). Auto-increments if taken. |\n| `--verbose` | Show all Next.js output including compilation details. |\n| `--reset` | Re-prompt for watch/output directories. |\n\n## Verbose mode\n\n```bash\nnpx doccupine --verbose\n```\n\n<Callout type=\"note\">\n  This will show Next.js output in the terminal, providing detailed logs useful for debugging during development.\n</Callout>\n\n## Generate the website\n\n```bash\nnpx doccupine build\n```\n\nThis scaffolds the Next.js app from your MDX files without starting the development server. To produce a production build, run `npm install && npm run build` in the generated output directory.\n\n## Show current configuration\n\n```bash\nnpx doccupine config --show\n```\n\nThis will show the current configuration for Doccupine.\n\n## Reset configuration\n\n```bash\nnpx doccupine config --reset\n```\n\nThis will reset the current configuration for Doccupine.";

package/dist/templates/mdx/commands.mdx.js

@@ -21,6 +21,14 @@ After selecting the directory, Doccupine will ask you to enter the name of the d
 
 This will start the development server on port 3000. Open your browser and navigate to http://localhost:3000 to view your documentation.
 
+## Options
+
+| Flag | Description |
+|---|---|
+| \`--port <port>\` | Port for the dev server (default: \`3000\`). Auto-increments if taken. |
+| \`--verbose\` | Show all Next.js output including compilation details. |
+| \`--reset\` | Re-prompt for watch/output directories. |
+
 ## Verbose mode
 
 \`\`\`bash
@@ -28,7 +36,7 @@ npx doccupine --verbose
 \`\`\`
 
 <Callout type="note">
-  This will show Next.js errors in the terminal console logs, providing more detailed output useful for debugging during development.
+  This will show Next.js output in the terminal, providing detailed logs useful for debugging during development.
 </Callout>
 
 ## Generate the website
@@ -37,7 +45,7 @@ npx doccupine --verbose
 npx doccupine build
 \`\`\`
 
-This will generate the build files for your documentation website without starting the development server. You can then deploy the generated files to a hosting service of your choice.
+This scaffolds the Next.js app from your MDX files without starting the development server. To produce a production build, run \`npm install && npm run build\` in the generated output directory.
 
 ## Show current configuration
 

package/dist/templates/mdx/components.mdx.d.ts

@@ -0,0 +1 @@
+export declare const componentsMdxTemplate = "---\ntitle: \"Components\"\ndescription: \"Explore the full library of built-in components available in your documentation pages.\"\ndate: \"2026-02-19\"\ncategory: \"Components\"\ncategoryOrder: 1\norder: 0\n---\n# Components\n\nDoccupine includes a rich set of built-in components you can use directly in your MDX files - no imports needed. Browse the full library below.\n\n<Columns cols={3}>\n  <Card title=\"Headers and Text\" icon=\"heading\" href=\"/headers-and-text\">\n    Headings, paragraphs, bold, italic, links, and other text formatting.\n  </Card>\n  <Card title=\"Lists and Tables\" icon=\"list\" href=\"/lists-and-tables\">\n    Ordered lists, unordered lists, and data tables.\n  </Card>\n  <Card title=\"Code\" icon=\"code\" href=\"/code\">\n    Inline code, fenced code blocks, and syntax highlighting.\n  </Card>\n  <Card title=\"Accordion\" icon=\"chevrons-down\" href=\"/accordion\">\n    Collapsible sections for organizing lengthy content.\n  </Card>\n  <Card title=\"Tabs\" icon=\"panel-top\" href=\"/tabs\">\n    Tabbed interfaces for grouping related content.\n  </Card>\n  <Card title=\"Cards\" icon=\"square\" href=\"/cards\">\n    Visual containers for content, icons, and links.\n  </Card>\n  <Card title=\"Buttons\" icon=\"mouse-pointer-click\" href=\"/buttons\">\n    Action components with variants, sizes, and icons.\n  </Card>\n  <Card title=\"Callouts\" icon=\"megaphone\" href=\"/callouts\">\n    Highlighted blocks for tips, warnings, and important notes.\n  </Card>\n  <Card title=\"Images and Embeds\" icon=\"image\" href=\"/images-and-embeds\">\n    Images, videos, and embedded content.\n  </Card>\n  <Card title=\"Icons\" icon=\"smile\" href=\"/icons\">\n    Lucide icons you can use anywhere in your docs.\n  </Card>\n  <Card title=\"Fields\" icon=\"text-cursor-input\" href=\"/fields\">\n    Property and parameter documentation blocks.\n  </Card>\n  <Card title=\"Update\" icon=\"bell\" href=\"/update\">\n    Changelog and update announcement blocks.\n  </Card>\n  <Card title=\"Columns\" icon=\"columns-3\" href=\"/columns\">\n    Grid layouts for arranging cards and content side by side.\n  </Card>\n  <Card title=\"Steps\" icon=\"footprints\" href=\"/steps\">\n    Numbered step-by-step guides and walkthroughs.\n  </Card>\n</Columns>";

package/dist/templates/mdx/components.mdx.js

@@ -0,0 +1,56 @@
+export const componentsMdxTemplate = `---
+title: "Components"
+description: "Explore the full library of built-in components available in your documentation pages."
+date: "2026-02-19"
+category: "Components"
+categoryOrder: 1
+order: 0
+---
+# Components
+
+Doccupine includes a rich set of built-in components you can use directly in your MDX files - no imports needed. Browse the full library below.
+
+<Columns cols={3}>
+  <Card title="Headers and Text" icon="heading" href="/headers-and-text">
+    Headings, paragraphs, bold, italic, links, and other text formatting.
+  </Card>
+  <Card title="Lists and Tables" icon="list" href="/lists-and-tables">
+    Ordered lists, unordered lists, and data tables.
+  </Card>
+  <Card title="Code" icon="code" href="/code">
+    Inline code, fenced code blocks, and syntax highlighting.
+  </Card>
+  <Card title="Accordion" icon="chevrons-down" href="/accordion">
+    Collapsible sections for organizing lengthy content.
+  </Card>
+  <Card title="Tabs" icon="panel-top" href="/tabs">
+    Tabbed interfaces for grouping related content.
+  </Card>
+  <Card title="Cards" icon="square" href="/cards">
+    Visual containers for content, icons, and links.
+  </Card>
+  <Card title="Buttons" icon="mouse-pointer-click" href="/buttons">
+    Action components with variants, sizes, and icons.
+  </Card>
+  <Card title="Callouts" icon="megaphone" href="/callouts">
+    Highlighted blocks for tips, warnings, and important notes.
+  </Card>
+  <Card title="Images and Embeds" icon="image" href="/images-and-embeds">
+    Images, videos, and embedded content.
+  </Card>
+  <Card title="Icons" icon="smile" href="/icons">
+    Lucide icons you can use anywhere in your docs.
+  </Card>
+  <Card title="Fields" icon="text-cursor-input" href="/fields">
+    Property and parameter documentation blocks.
+  </Card>
+  <Card title="Update" icon="bell" href="/update">
+    Changelog and update announcement blocks.
+  </Card>
+  <Card title="Columns" icon="columns-3" href="/columns">
+    Grid layouts for arranging cards and content side by side.
+  </Card>
+  <Card title="Steps" icon="footprints" href="/steps">
+    Numbered step-by-step guides and walkthroughs.
+  </Card>
+</Columns>`;

package/dist/templates/mdx/deployment.mdx.d.ts

@@ -1 +1 @@
-export declare const deploymentMdxTemplate = "---\ntitle: \"Deployment\"\ndescription: \"Deploy your documentation site with the Doccupine Platform or self-host on any platform that supports Next.js.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 10\n---\n# Deployment\nThe fastest way to deploy your documentation is with the [Doccupine Platform](https://www.doccupine.com). If you prefer to manage your own infrastructure, you can self-host the generated Next.js app on any platform.\n\n## Doccupine Platform\n\nSign up at [doccupine.com](https://www.doccupine.com) and connect your repository. Your documentation site is live in minutes - no build configuration, no infrastructure to manage.\n\n<Callout type=\"success\">\n  The Doccupine Platform is the recommended way to deploy. It handles builds, hosting, SSL, and updates automatically so you can focus on writing documentation.\n</Callout>\n\n### What you get\n- **Automatic deployments** on every push to your repository\n- **Site customization** through a visual dashboard - no code changes needed\n- **Team collaboration** so your whole team can manage docs together\n- **Custom domains** with automatic SSL\n- **AI Assistant and MCP server** included out of the box, no API key required\n- **Zero maintenance** - no servers, no build pipelines, no dependency updates\n\n### Getting started\n1. Create an account at [doccupine.com](https://www.doccupine.com).\n2. Connect your GitHub repository.\n3. Your site is deployed automatically.\n\nEvery push to your repository triggers a new deployment. You can customize your site's appearance, domain, and settings from the dashboard.\n\n---\n\n## Self-hosting\n\nDoccupine generates a standard Next.js app, so you can deploy it anywhere that supports Node.js or Next.js.\n\n<Callout type=\"warning\">\n  Deploy the generated website directory (the Next.js app), not your MDX source folder. In a monorepo, set the root directory to the generated site folder.\n</Callout>\n\n<Callout type=\"note\">\n  Self-hosting requires you to manage your own build pipeline, hosting, SSL certificates, and AI provider API keys. For a hands-off experience, consider the [Doccupine Platform](https://www.doccupine.com).\n</Callout>\n\n### Popular hosting options\n\n- **Vercel** - native Next.js support, zero-config deploys. Connect your repo and set the root directory to the generated app folder.\n- **Netlify** - supports Next.js via the `@netlify/plugin-nextjs` adapter. Works with the standard `next build` output.\n- **AWS Amplify** - fully managed hosting with CI/CD. Supports Next.js SSR out of the box.\n- **Cloudflare Pages** - deploy using the `@cloudflare/next-on-pages` adapter for edge-based hosting.\n- **Docker** - build a container from the generated app using the standard [Next.js Docker example](https://github.com/vercel/next.js/tree/canary/examples/with-docker) and deploy to any container platform.\n- **Node.js server** - run `next build && next start` on any server or VPS with Node.js installed.\n\n### Troubleshooting\n- **Build failed** - check build logs. Ensure your lockfile and correct Node.js version are present.\n- **Missing content** - verify your MDX files and assets are in the repository.\n- **SSR issues on edge platforms** - some features (like the AI chat API routes) require a Node.js runtime. Check your platform's documentation for SSR/API route support.";
+export declare const deploymentMdxTemplate = "---\ntitle: \"Deployment\"\ndescription: \"Deploy your documentation site with the Doccupine Platform or self-host on any platform that supports Next.js.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 10\n---\n# Deployment\nThe fastest way to deploy your documentation is with the [Doccupine Platform](https://www.doccupine.com). If you prefer to manage your own infrastructure, you can self-host the generated Next.js app on any platform.\n\n## Doccupine Platform\n\nSign up at [doccupine.com](https://www.doccupine.com) and connect your repository. Your documentation site is live in minutes - no build configuration, no infrastructure to manage.\n\n<Callout type=\"success\">\n  The Doccupine Platform is the recommended way to deploy. It handles builds, hosting, SSL, and updates automatically so you can focus on writing documentation.\n</Callout>\n\n### What you get\n- **Automatic deployments** on every push to your repository\n- **Site customization** through a visual dashboard - no code changes needed\n- **Team collaboration** so your whole team can manage docs together\n- **Custom domains** with automatic SSL\n- **AI Assistant and MCP server** included out of the box, no API key required\n- **Zero maintenance** - no servers, no build pipelines, no dependency updates\n\n### Getting started\n1. Create an account at [doccupine.com](https://www.doccupine.com).\n2. Connect your GitHub repository.\n3. Your site is deployed automatically.\n\nEvery push to your repository triggers a new deployment. You can customize your site's appearance, domain, and settings from the dashboard. See the [Platform Overview](/platform) for a full walkthrough of the dashboard, editor, and configuration options.\n\n---\n\n## Self-hosting\n\nDoccupine generates a standard Next.js app, so you can deploy it anywhere that supports Node.js or Next.js.\n\n<Callout type=\"warning\">\n  Deploy the generated website directory (the Next.js app), not your MDX source folder. In a monorepo, set the root directory to the generated site folder.\n</Callout>\n\n<Callout type=\"note\">\n  Self-hosting requires you to manage your own build pipeline, hosting, SSL certificates, and AI provider API keys. For a hands-off experience, consider the [Doccupine Platform](https://www.doccupine.com).\n</Callout>\n\n### Popular hosting options\n\n- **Vercel** - native Next.js support, zero-config deploys. Connect your repo and set the root directory to the generated app folder.\n- **Netlify** - supports Next.js via the `@netlify/plugin-nextjs` adapter. Works with the standard `next build` output.\n- **AWS Amplify** - fully managed hosting with CI/CD. Supports Next.js SSR out of the box.\n- **Cloudflare Pages** - deploy using the `@cloudflare/next-on-pages` adapter for edge-based hosting.\n- **Docker** - build a container from the generated app using the standard [Next.js Docker example](https://github.com/vercel/next.js/tree/canary/examples/with-docker) and deploy to any container platform.\n- **Node.js server** - run `next build && next start` on any server or VPS with Node.js installed.\n\n### Troubleshooting\n- **Build failed** - check build logs. Ensure your lockfile and correct Node.js version are present.\n- **Missing content** - verify your MDX files and assets are in the repository.\n- **SSR issues on edge platforms** - some features (like the AI chat API routes) require a Node.js runtime. Check your platform's documentation for SSR/API route support.";

package/dist/templates/mdx/deployment.mdx.js

@@ -30,7 +30,7 @@ Sign up at [doccupine.com](https://www.doccupine.com) and connect your repositor
 2. Connect your GitHub repository.
 3. Your site is deployed automatically.
 
-Every push to your repository triggers a new deployment. You can customize your site's appearance, domain, and settings from the dashboard.
+Every push to your repository triggers a new deployment. You can customize your site's appearance, domain, and settings from the dashboard. See the [Platform Overview](/platform) for a full walkthrough of the dashboard, editor, and configuration options.
 
 ---
 

package/dist/templates/mdx/globals.mdx.d.ts

@@ -1 +1 @@
-export declare const globalsMdxTemplate = "---\ntitle: \"Globals\"\ndescription: \"Configure global settings for your documentation.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 1\n---\n# Global Configuration\nUse a `config.json` file to define project\u2011wide metadata for your documentation site. These values are applied to every generated page unless a page overrides them in its own frontmatter.\n\n## config.json\nPlace a `config.json` at your project root (the same folder where you execute `npx doccupine`) to define global metadata for your documentation site.\n\n```json\n{\n  \"name\": \"Doccupine\",\n  \"description\": \"Doccupine is a free and open-source documentation platform. Write MDX, get a production-ready site with AI chat, built-in components, and an MCP server - in one command.\",\n  \"icon\": \"https://docs.doccupine.com/favicon.ico\",\n  \"image\": \"https://docs.doccupine.com/preview.png\"\n}\n```\n\n## Fields\n- **name**: The primary name of your documentation website. Displayed in the site title and used in various UI elements.\n- **description**: A concise summary of your project, used in site metadata (e.g., HTML meta description) and social previews when not overridden.\n- **icon**: The favicon for your site. You can provide a full URL or a relative path to an asset in your project.\n- **image**: The Open Graph image used when links to your docs are shared on social platforms. Accepts a full URL or a relative path.\n\n## Per-page overrides\nAny page can override global values by defining the matching key in its frontmatter. When present, the page's value takes precedence over `config.json` for that page only.\n\n| Frontmatter field | Overrides | Effect |\n|---|---|---|\n| **title** | - | Page title in metadata and Open Graph |\n| **description** | `description` | Meta description and Open Graph description |\n| **name** | `name` | Site name shown in the title suffix (e.g. \"Page - My Docs\") |\n| **icon** | `icon` | Favicon for this page |\n| **image** | `image` | Open Graph preview image |\n\n<Callout type=\"note\">\n  If a key is not specified in a page's frontmatter, Doccupine falls back to the corresponding value in `config.json`.\n</Callout>\n\nExample frontmatter in an `.mdx` file:\n\n```text\n---\ntitle: \"My Feature\"\ndescription: \"A focused description just for this page.\"\nname: \"My Product Docs\"\nicon: \"/custom-favicon.ico\"\nimage: \"/custom-preview.png\"\ndate: \"2026-02-19\"\ncategory: \"Guides\"\n---\n```\n\n";
+export declare const globalsMdxTemplate = "---\ntitle: \"Globals\"\ndescription: \"Configure global settings for your documentation.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 1\n---\n# Global Configuration\nUse a `config.json` file to define project\u2011wide metadata for your documentation site. These values are applied to every generated page unless a page overrides them in its own frontmatter.\n\n## config.json\nPlace a `config.json` at your project root (the same folder where you execute `npx doccupine`) to define global metadata for your documentation site.\n\n```json\n{\n  \"name\": \"Doccupine\",\n  \"description\": \"Doccupine is a free and open-source documentation platform. Write MDX, get a production-ready site with AI chat, built-in components, and an MCP server - in one command.\",\n  \"icon\": \"https://docs.doccupine.com/favicon.ico\",\n  \"image\": \"https://docs.doccupine.com/preview.png\"\n}\n```\n\n## Fields\nAll fields are optional. Doccupine uses sensible defaults when a field is not set.\n\n- **name**: The primary name of your documentation website. Displayed in the site title and used in various UI elements.\n- **description**: A concise summary of your project, used in site metadata (e.g., HTML meta description) and social previews when not overridden.\n- **icon**: The favicon for your site. You can provide a full URL or a relative path to an asset in your project.\n- **image**: The Open Graph image used when links to your docs are shared on social platforms. Accepts a full URL or a relative path.\n\n## Per-page overrides\nAny page can override global values by defining the matching key in its frontmatter. When present, the page's value takes precedence over `config.json` for that page only.\n\n| Frontmatter field | Overrides | Effect |\n|---|---|---|\n| **title** | - | Page title in metadata and Open Graph |\n| **description** | `description` | Meta description and Open Graph description |\n| **name** | `name` | Site name shown in the title suffix (e.g. \"Page - My Docs\") |\n| **icon** | `icon` | Favicon for this page |\n| **image** | `image` | Open Graph preview image |\n| **section** | - | Assigns the page to a [section](/sections) |\n| **sectionOrder** | - | Controls section position in the tab bar |\n| **sectionLabel** | - | Renames the default \"Docs\" tab (use on `index.mdx`) |\n\n<Callout type=\"note\">\n  If a key is not specified in a page's frontmatter, Doccupine falls back to the corresponding value in `config.json`.\n</Callout>\n\nExample frontmatter in an `.mdx` file:\n\n```text\n---\ntitle: \"My Feature\"\ndescription: \"A focused description just for this page.\"\nname: \"My Product Docs\"\nicon: \"/custom-favicon.ico\"\nimage: \"/custom-preview.png\"\ndate: \"2026-02-19\"\ncategory: \"Guides\"\n---\n```\n\n";

package/dist/templates/mdx/globals.mdx.js

@@ -23,6 +23,8 @@ Place a \`config.json\` at your project root (the same folder where you execute
 \`\`\`
 
 ## Fields
+All fields are optional. Doccupine uses sensible defaults when a field is not set.
+
 - **name**: The primary name of your documentation website. Displayed in the site title and used in various UI elements.
 - **description**: A concise summary of your project, used in site metadata (e.g., HTML meta description) and social previews when not overridden.
 - **icon**: The favicon for your site. You can provide a full URL or a relative path to an asset in your project.
@@ -38,6 +40,9 @@ Any page can override global values by defining the matching key in its frontmat
 | **name** | \`name\` | Site name shown in the title suffix (e.g. "Page - My Docs") |
 | **icon** | \`icon\` | Favicon for this page |
 | **image** | \`image\` | Open Graph preview image |
+| **section** | - | Assigns the page to a [section](/sections) |
+| **sectionOrder** | - | Controls section position in the tab bar |
+| **sectionLabel** | - | Renames the default "Docs" tab (use on \`index.mdx\`) |
 
 <Callout type="note">
   If a key is not specified in a page's frontmatter, Doccupine falls back to the corresponding value in \`config.json\`.

package/dist/templates/mdx/index.mdx.d.ts

@@ -1 +1 @@
-export declare const indexMdxTemplate = "---\ntitle: \"Introduction\"\ndescription: \"Doccupine is a free and open-source documentation platform. Write MDX, get a production-ready site with AI chat, built-in components, and an MCP server - in one command.\"\ndate: \"2026-02-19\"\ncategory: \"Getting Started\"\ncategoryOrder: 0\norder: 0\n---\n# Welcome to Doccupine\n\nDoccupine turns a folder of MDX files into a beautiful, production-ready documentation website. Write in standard Markdown, use any of the built-in components, and get a fully themed site with AI-powered search and automatic navigation - all from one command.\n\n```bash\nnpx doccupine\n```\n\nThat's it. Doccupine prompts you for a source directory, generates a Next.js site, and starts a dev server at `http://localhost:3000`. Every edit rebuilds instantly.\n\n<Callout type=\"info\">\n  Want a managed experience instead? The [Doccupine Platform](/platform) gives you a browser-based editor, one-click publishing, custom domains, and team collaboration - no local setup needed.\n</Callout>\n\n## How it works\n\nYou write `.mdx` files using Markdown and built-in components like this one. Doccupine watches your files, generates pages, builds the sidebar from your frontmatter, and handles theming, AI chat, and deployment config for you.\n\nEverything on this page - the callout above, the cards below, the code block - is a live example of what your docs can look like out of the box.\n\n<Columns cols={2}>\n  <Card title=\"Start writing\" icon=\"pencil\">\n    Rich docs from day one. [14+ components](/headers-and-text) work out of the box - no imports needed.\n  </Card>\n  <Card title=\"Ship it\" icon=\"rocket\">\n    Custom domains and team collaboration on the [Doccupine Platform](/platform), or [self-host](/deployment) anywhere.\n  </Card>\n</Columns>\n\n## Configure your site\n\nDoccupine works with no configuration, but when you're ready to customize, everything is controlled through simple JSON files in your project root.\n\n- [Global settings](/globals) - Site name, description, favicon, and preview image via `config.json`\n- [Navigation](/navigation) - Override the auto-generated sidebar with `navigation.json`\n- [Theme](/theme) - Colors, logos, and dark mode palette via `theme.json`\n- [Fonts](/fonts) - Google Fonts or local font files via `fonts.json`\n- [AI Assistant](/ai-assistant) - Enable the built-in chat with your own API key\n- [MCP Server](/model-context-protocol) - Let AI tools search your docs through `/api/mcp`\n";
+export declare const indexMdxTemplate = "---\ntitle: \"Introduction\"\ndescription: \"Doccupine is a free and open-source documentation platform. Write MDX, get a production-ready site with AI chat, built-in components, and an MCP server - in one command.\"\ndate: \"2026-02-19\"\ncategory: \"Getting Started\"\ncategoryOrder: 0\norder: 0\n---\n# Welcome to Doccupine\n\nDoccupine turns a folder of MDX files into a beautiful, production-ready documentation website. Write in standard Markdown, use any of the built-in components, and get a fully themed site with AI-powered search and automatic navigation - all from one command.\n\n```bash\nnpx doccupine\n```\n\nThat's it. Doccupine prompts you for a source directory, generates your site, and starts a dev server at `http://localhost:3000`. Every edit rebuilds instantly.\n\n<Callout type=\"info\">\n  Want a managed experience instead? The [Doccupine Platform](/platform) gives you a browser-based editor, one-click publishing, custom domains, and team collaboration - no local setup needed.\n</Callout>\n\n## How it works\n\nYou write `.mdx` files using Markdown and built-in components like this one. Doccupine watches your files, generates pages, builds the sidebar from your frontmatter, and handles theming, AI chat, and deployment config for you.\n\nEverything on this page - the callout above, the cards below, the code block - is a live example of what your docs can look like out of the box.\n\n<Columns cols={2}>\n  <Card title=\"Start writing\" icon=\"pencil\" href=\"/components\">\n    Rich docs from day one. 14+ components work out of the box - no imports needed.\n  </Card>\n  <Card title=\"Ship it\" icon=\"rocket\" href=\"/deployment\">\n    Deploy to the Doccupine Platform or self-host on any platform that supports Next.js.\n  </Card>\n</Columns>\n\n## Configure your site\n\nDoccupine works with no configuration, but when you're ready to customize, everything is controlled through simple JSON files in your project root.\n\n- [Global settings](/globals) - Site name, description, favicon, and preview image via `config.json`\n- [Navigation](/navigation) - Override the auto-generated sidebar with `navigation.json`\n- [Theme](/theme) - Colors, logos, and dark mode palette via `theme.json`\n- [Fonts](/fonts) - Google Fonts or local font files via `fonts.json`\n- [AI Assistant](/ai-assistant) - Enable the built-in chat with your own API key\n- [MCP Server](/model-context-protocol) - Let AI tools search your docs through `/api/mcp`\n";

package/dist/templates/mdx/index.mdx.js

@@ -15,7 +15,7 @@ Doccupine turns a folder of MDX files into a beautiful, production-ready documen
 npx doccupine
 \`\`\`
 
-That's it. Doccupine prompts you for a source directory, generates a Next.js site, and starts a dev server at \`http://localhost:3000\`. Every edit rebuilds instantly.
+That's it. Doccupine prompts you for a source directory, generates your site, and starts a dev server at \`http://localhost:3000\`. Every edit rebuilds instantly.
 
 <Callout type="info">
   Want a managed experience instead? The [Doccupine Platform](/platform) gives you a browser-based editor, one-click publishing, custom domains, and team collaboration - no local setup needed.
@@ -28,11 +28,11 @@ You write \`.mdx\` files using Markdown and built-in components like this one. D
 Everything on this page - the callout above, the cards below, the code block - is a live example of what your docs can look like out of the box.
 
 <Columns cols={2}>
-  <Card title="Start writing" icon="pencil">
-    Rich docs from day one. [14+ components](/headers-and-text) work out of the box - no imports needed.
+  <Card title="Start writing" icon="pencil" href="/components">
+    Rich docs from day one. 14+ components work out of the box - no imports needed.
   </Card>
-  <Card title="Ship it" icon="rocket">
-    Custom domains and team collaboration on the [Doccupine Platform](/platform), or [self-host](/deployment) anywhere.
+  <Card title="Ship it" icon="rocket" href="/deployment">
+    Deploy to the Doccupine Platform or self-host on any platform that supports Next.js.
   </Card>
 </Columns>
 

package/dist/templates/mdx/model-context-protocol.mdx.d.ts

@@ -1 +1 @@
-export declare const mcpMdxTemplate = "---\ntitle: \"Model Context Protocol\"\ndescription: \"Connect your Doccupine documentation to AI tools with an MCP server for enhanced AI-powered documentation search.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 9\n---\n# Model Context Protocol\nConnect your documentation to AI tools with a hosted MCP server.\n\nDoccupine automatically generates a Model Context Protocol (MCP) server from your documentation, making your content accessible to AI applications like Claude, Cursor, VS Code, and other MCP-compatible tools. Your MCP server exposes semantic search capabilities, allowing AI tools to query your documentation directly and provide accurate, context-aware answers.\n\n<Callout type=\"warning\">\n  The MCP server requires the [AI Assistant](/ai-assistant) to be configured. Make sure you have set up your LLM provider and API keys before using the MCP server.\n</Callout>\n\n## About MCP servers\n\nThe Model Context Protocol (MCP) is an open protocol that creates standardized connections between AI applications and external services, like documentation. Doccupine generates an MCP server from your documentation, preparing your content for the broader AI ecosystem where any MCP client can connect to your documentation.\n\nYour MCP server exposes search and retrieval tools for AI applications to query your documentation. Your users must connect your MCP server to their preferred AI tools to access your documentation.\n\n### How MCP servers work\n\nWhen an AI tool has your documentation MCP server connected, the AI tool can search your documentation directly instead of making a generic web search in response to a user's prompt. Your MCP server provides access to all indexed content from your documentation site.\n\n* The LLM can proactively search your documentation while generating a response, not just when explicitly asked.\n* The LLM determines when to use the search tool based on the context of the conversation and the relevance of your documentation.\n* Each tool call happens during the generation process, so the LLM searches up-to-date information from your documentation to generate its response.\n\n### MCP compared to web search\n\nAI tools can search the web, but MCP provides distinct advantages for documentation.\n\n* **Direct source access**: Web search depends on what search engines have indexed, which may be stale or incomplete. MCP searches your current indexed documentation directly.\n* **Integrated workflow**: MCP allows the AI to search during response generation rather than performing a separate web search.\n* **Semantic search**: MCP uses vector embeddings for semantic similarity search, providing more relevant results than keyword-based web search.\n* **No search noise**: SEO and ranking algorithms influence web search results. MCP goes straight to your documentation content.\n\n## Access your MCP server\n\nDoccupine automatically generates an MCP server for your documentation and hosts it at your documentation URL with the `/api/mcp` path. For example, if your documentation is hosted at `https://example.com`, your MCP server is available at `https://example.com/api/mcp`.\n\nThe MCP server provides both a GET endpoint to discover available tools and a POST endpoint to execute tool calls.\n\n### Authentication\n\nYou can optionally protect your MCP server with an API key by setting the `DOCS_API_KEY` environment variable in your `.env` file:\n\n```bash\nDOCS_API_KEY=your-secret-key\n```\n\nWhen `DOCS_API_KEY` is set, all requests to `/api/mcp` must include an `Authorization` header with a Bearer token:\n\n```text\nAuthorization: Bearer your-secret-key\n```\n\nRequests without a valid token receive a `401 Unauthorized` response. When `DOCS_API_KEY` is not set, the MCP server is publicly accessible with no authentication required.\n\n<Callout type=\"note\">\n  Authentication only applies to the `/api/mcp` endpoint. The `/api/rag` endpoint used by the built-in AI Assistant chat is not affected by this setting.\n</Callout>\n\n### API Endpoints\n\n#### GET /api/mcp\n\nReturns information about available tools and the current index status.\n\n**Response:**\n```json\n{\n  \"tools\": [\n    {\n      \"name\": \"search_docs\",\n      \"description\": \"Search through the documentation content using semantic search...\",\n      \"inputSchema\": { ... }\n    },\n    ...\n  ],\n  \"index\": {\n    \"ready\": true,\n    \"chunkCount\": 150\n  }\n}\n```\n\n#### POST /api/mcp\n\nExecutes an MCP tool call.\n\n**Request Body:**\n```json\n{\n  \"tool\": \"search_docs\",\n  \"params\": {\n    \"query\": \"how to deploy\",\n    \"limit\": 6\n  }\n}\n```\n\n**Response:**\n```json\n{\n  \"content\": [\n    {\n      \"path\": \"app/deployment/page.tsx\",\n      \"uri\": \"docs://deployment\",\n      \"score\": \"0.892\",\n      \"text\": \"Deploy your Doccupine site as a Next.js application...\"\n    },\n    ...\n  ]\n}\n```\n\n## Available Tools\n\nYour MCP server exposes three tools for interacting with your documentation:\n\n### search_docs\n\nSearch through the documentation content using semantic search. Returns relevant chunks of documentation based on the query using vector embeddings and cosine similarity.\n\n**Parameters:**\n- `query` (required): The search query to find relevant documentation\n- `limit` (optional): Maximum number of results to return (default: 6)\n\n**Example:**\n```json\n{\n  \"tool\": \"search_docs\",\n  \"params\": {\n    \"query\": \"how to configure AI assistant\",\n    \"limit\": 5\n  }\n}\n```\n\n### get_doc\n\nGet the full content of a specific documentation page by its path.\n\n**Parameters:**\n- `path` (required): The file path to the documentation page (e.g., `app/getting-started/page.tsx`)\n\n**Example:**\n```json\n{\n  \"tool\": \"get_doc\",\n  \"params\": {\n    \"path\": \"app/deployment/page.tsx\"\n  }\n}\n```\n\n### list_docs\n\nList all available documentation pages, optionally filtered by directory.\n\n**Parameters:**\n- `directory` (optional): Optional directory to filter results (e.g., `components`)\n\n**Example:**\n```json\n{\n  \"tool\": \"list_docs\",\n  \"params\": {\n    \"directory\": \"configuration\"\n  }\n}\n```\n\n## How it works\n\nDoccupine's MCP server uses semantic search powered by vector embeddings to provide accurate, context-aware search results.\n\n### Indexing Process\n\n1. **Document Discovery**: The server scans your `app/` directory for all `page.tsx`, `page.ts`, `page.jsx`, and `page.js` files.\n2. **Content Extraction**: It extracts content from `const content =` declarations in your page files.\n3. **Chunking**: Large documents are split into chunks of approximately 1500 characters with 200 character overlap for better context preservation.\n4. **Embedding Generation**: Each chunk is converted to a vector embedding using your configured LLM provider's embedding model.\n5. **Index Building**: All embeddings are stored in memory for fast similarity search.\n\n### Search Process\n\n1. **Query Embedding**: The search query is converted to a vector embedding using the same embedding model.\n2. **Similarity Calculation**: Cosine similarity is calculated between the query embedding and all document chunk embeddings.\n3. **Ranking**: Results are sorted by similarity score (highest first).\n4. **Response**: The top N results (based on the limit parameter) are returned with their paths, URIs, scores, and text content.\n\n<Callout type=\"note\">\n  The index is built automatically when the first search is performed. The index is stored in memory and persists for the lifetime of the server process. If you update your documentation, restart the server to rebuild the index.\n</Callout>\n\n## Use your MCP server\n\nYour users must connect your MCP server to their preferred AI tools.\n\n1. Make your MCP server URL publicly available.\n2. Users copy your MCP server URL and add it to their tools.\n3. Users access your documentation through their tools.\n\nThese are some of the ways you can help your users connect to your MCP server:\n\n<Tabs>\n  <TabContent title=\"Claude\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the steps to connect it to Claude.\n        1. Navigate to the [Connectors](https://claude.ai/settings/connectors) page in the Claude settings.\n        2. Select **Add custom connector**.\n        3. Add your MCP server name and URL.\n        4. Select **Add**.\n        5. When using Claude, select the attachments button (the plus icon).\n        6. Select your MCP server.\n      </Step>\n    </Steps>\n    See the [Model Context Protocol documentation](https://modelcontextprotocol.io/docs/tutorials/use-remote-mcp-server#connecting-to-a-remote-mcp-server) for more details.\n  </TabContent>\n  <TabContent title=\"Claude Code\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the command to connect it to Claude Code.\n        ```bash\n        claude mcp add --transport http <name> <url>\n        ```\n      </Step>\n    </Steps>\n    See the [Claude Code documentation](https://docs.anthropic.com/en/docs/claude-code/mcp#installing-mcp-servers) for more details.\n  </TabContent>\n  <TabContent title=\"Cursor\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the steps to connect it to Cursor.\n        1. Use <kbd>Command</kbd> + <kbd>Shift</kbd> + <kbd>P</kbd> (<kbd>Ctrl</kbd> + <kbd>Shift</kbd> + <kbd>P</kbd> on Windows) to open the command palette.\n        2. Search for \"Open MCP settings\".\n        3. Select **Add custom MCP**. This opens the `mcp.json` file.\n        4. In `mcp.json`, configure your server:\n        ```json\n        {\n          \"mcpServers\": {\n            \"<your-mcp-server-name>\": {\n              \"url\": \"<your-mcp-server-url>\"\n            }\n          }\n        }\n        ```\n      </Step>\n    </Steps>\n    See the [Cursor documentation](https://docs.cursor.com/en/context/mcp#installing-mcp-servers) for more details.\n  </TabContent>\n  <TabContent title=\"VS Code\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the steps to connect it to VS Code.\n        1. Create a `.vscode/mcp.json` file.\n        2. In `mcp.json`, configure your server:\n        ```json\n        {\n          \"servers\": {\n            \"<your-mcp-server-name>\": {\n              \"type\": \"http\",\n              \"url\": \"<your-mcp-server-url>\"\n            }\n          }\n        }\n        ```\n      </Step>\n    </Steps>\n    See the [VS Code documentation](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) for more details.\n  </TabContent>\n</Tabs>\n\n## Requirements\n\nTo use the MCP server, you need to have the AI Assistant configured. The MCP server uses the same LLM configuration for generating embeddings.\n\n| Variable | Required | Description |\n|---|---|---|\n| `LLM_PROVIDER` | Yes | Your LLM provider (`openai`, `anthropic`, or `google`) |\n| `DOCS_API_KEY` | No | When set, requires Bearer token authentication on `/api/mcp` |\n\n<Callout type=\"warning\">\n  The MCP server requires an LLM provider to be configured for generating embeddings. Make sure you have set up your AI Assistant with a valid API key before using the MCP server.\n</Callout>\n\nSee the [AI Assistant documentation](/ai-assistant) for configuration details.\n\n## Content indexing\n\nYour MCP server searches content extracted from your page files. The server automatically discovers and indexes all `page.tsx`, `page.ts`, `page.jsx`, and `page.js` files in your `app/` directory.\n\n### Content extraction\n\nThe server extracts content from `const content =` declarations in your page files. Make sure your documentation pages export a `content` constant with your markdown or MDX content.\n\n**Example:**\n```tsx\nexport const content = `\n# Getting Started\n\nWelcome to the documentation...\n`;\n```\n\n### Excluded directories\n\nThe following directories are automatically excluded from indexing:\n- `node_modules`\n- `.next`\n- `.git`\n- `api`\n\n## Troubleshooting\n\n### Index not building\n\nIf the index is not building, check:\n- Your LLM provider is configured correctly in your `.env` file\n- You have a valid API key set\n- Your documentation pages export a `content` constant\n\n### No search results\n\nIf searches return no results:\n- Verify that your documentation pages are in the `app/` directory\n- Check that your pages export a `content` constant\n- Ensure the index has been built (check the `index.ready` status via GET `/api/mcp`)\n\n### Slow search performance\n\nThe first search may be slower as it builds the index. Subsequent searches are fast as they use the in-memory index. If performance is consistently slow:\n- Check your embedding API response times\n- Consider reducing the number of documentation pages\n- Verify your server has sufficient memory\n\n### Cloudflare blocking MCP requests\n\nIf you use Cloudflare as a proxy in front of your documentation site, Cloudflare's bot protection may block server-to-server requests from AI tools like Claude.ai. This can cause MCP connections to fail silently or return errors.\n\nThere are two ways to resolve this:\n\n**Option 1: Disable the Cloudflare proxy (simplest)**\n\nIn your Cloudflare DNS settings, click the orange cloud icon next to your domain record to switch it to \"DNS only\" (grey cloud). This disables Cloudflare's proxy and bot protection for your domain, allowing MCP requests to reach your server directly.\n\n**Option 2: Add a Cloudflare WAF exception (keeps your custom domain proxied)**\n\nIn Cloudflare dashboard:\n\n1. Go to **Security > WAF**.\n2. Click **Create rule**.\n3. Set it up as:\n   - **Rule name**: Allow MCP API\n   - **Field**: URI Path\n   - **Operator**: starts with\n   - **Value**: `/api/mcp`\n   - **Action**: Skip -- then check all remaining custom rules, Rate limiting rules, and Bot Fight Mode / Super Bot Fight Mode.\n4. Deploy the rule and make sure it is ordered first (above other rules).\n\n<Callout type=\"warning\">\n  Also check **Security > Bots** in your Cloudflare dashboard. If \"Bot Fight Mode\" or \"Super Bot Fight Mode\" is enabled, that is very likely what is blocking server-to-server requests from AI tools.\n</Callout>\n\n## Best practices\n\n* **Keep content up-to-date**: Restart your server after updating documentation to rebuild the index with fresh content.\n* **Use descriptive queries**: More specific queries yield better semantic search results.\n* **Monitor index status**: Use the GET endpoint to check if your index is ready before performing searches.\n* **Optimize content structure**: Well-structured markdown with clear headings improves search relevance.";
+export declare const mcpMdxTemplate = "---\ntitle: \"Model Context Protocol\"\ndescription: \"Connect your Doccupine documentation to AI tools with an MCP server for enhanced AI-powered documentation search.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 9\n---\n# Model Context Protocol\nConnect your documentation to AI tools with a hosted MCP server.\n\nDoccupine automatically generates a Model Context Protocol (MCP) server from your documentation, making your content accessible to AI applications like Claude, Cursor, VS Code, and other MCP-compatible tools. Your MCP server exposes semantic search capabilities, allowing AI tools to query your documentation directly and provide accurate, context-aware answers.\n\n<Callout type=\"warning\">\n  The MCP server requires the [AI Assistant](/ai-assistant) to be configured. Make sure you have set up your LLM provider and API keys before using the MCP server.\n</Callout>\n\n## About MCP servers\n\nThe Model Context Protocol (MCP) is an open protocol that creates standardized connections between AI applications and external services, like documentation. Doccupine generates an MCP server from your documentation, preparing your content for the broader AI ecosystem where any MCP client can connect to your documentation.\n\nYour MCP server exposes search and retrieval tools for AI applications to query your documentation. Your users must connect your MCP server to their preferred AI tools to access your documentation.\n\n### How MCP servers work\n\nWhen an AI tool has your documentation MCP server connected, the AI tool can search your documentation directly instead of making a generic web search in response to a user's prompt. Your MCP server provides access to all indexed content from your documentation site.\n\n* The LLM can proactively search your documentation while generating a response, not just when explicitly asked.\n* The LLM determines when to use the search tool based on the context of the conversation and the relevance of your documentation.\n* Each tool call happens during the generation process, so the LLM searches up-to-date information from your documentation to generate its response.\n\n### MCP compared to web search\n\nAI tools can search the web, but MCP provides distinct advantages for documentation.\n\n* **Direct source access**: Web search depends on what search engines have indexed, which may be stale or incomplete. MCP searches your current indexed documentation directly.\n* **Integrated workflow**: MCP allows the AI to search during response generation rather than performing a separate web search.\n* **Semantic search**: MCP uses vector embeddings for semantic similarity search, providing more relevant results than keyword-based web search.\n* **No search noise**: SEO and ranking algorithms influence web search results. MCP goes straight to your documentation content.\n\n## Access your MCP server\n\nDoccupine automatically generates an MCP server for your documentation and hosts it at your documentation URL with the `/api/mcp` path. For example, if your documentation is hosted at `https://example.com`, your MCP server is available at `https://example.com/api/mcp`.\n\nThe MCP server provides both a GET endpoint to discover available tools and a POST endpoint to execute tool calls.\n\n### Authentication\n\nYou can optionally protect your MCP server with an API key by setting the `DOCS_API_KEY` environment variable in your `.env` file:\n\n```bash\nDOCS_API_KEY=your-secret-key\n```\n\nWhen `DOCS_API_KEY` is set, all requests to `/api/mcp` must include an `Authorization` header with a Bearer token:\n\n```text\nAuthorization: Bearer your-secret-key\n```\n\nRequests without a valid token receive a `401 Unauthorized` response. When `DOCS_API_KEY` is not set, the MCP server is publicly accessible with no authentication required.\n\n<Callout type=\"note\">\n  Authentication only applies to the `/api/mcp` endpoint. The `/api/rag` endpoint used by the built-in AI Assistant chat is not affected by this setting.\n</Callout>\n\n### API Endpoints\n\n#### GET /api/mcp\n\nReturns information about available tools and the current index status.\n\n**Response:**\n```json\n{\n  \"tools\": [\n    {\n      \"name\": \"search_docs\",\n      \"description\": \"Search through the documentation content using semantic search...\",\n      \"inputSchema\": { ... }\n    },\n    ...\n  ],\n  \"index\": {\n    \"ready\": true,\n    \"chunkCount\": 150\n  }\n}\n```\n\n#### POST /api/mcp\n\nExecutes an MCP tool call.\n\n**Request Body:**\n```json\n{\n  \"tool\": \"search_docs\",\n  \"params\": {\n    \"query\": \"how to deploy\",\n    \"limit\": 6\n  }\n}\n```\n\n**Response:**\n```json\n{\n  \"content\": [\n    {\n      \"path\": \"app/deployment/page.tsx\",\n      \"uri\": \"docs://deployment\",\n      \"score\": \"0.892\",\n      \"text\": \"Deploy your Doccupine site as a Next.js application...\"\n    },\n    ...\n  ]\n}\n```\n\n## Available Tools\n\nYour MCP server exposes three tools for interacting with your documentation:\n\n### search_docs\n\nSearch through the documentation content using semantic search. Returns relevant chunks of documentation based on the query using vector embeddings and cosine similarity.\n\n**Parameters:**\n- `query` (required): The search query to find relevant documentation\n- `limit` (optional): Maximum number of results to return (default: 6)\n\n**Example:**\n```json\n{\n  \"tool\": \"search_docs\",\n  \"params\": {\n    \"query\": \"how to configure AI assistant\",\n    \"limit\": 5\n  }\n}\n```\n\n### get_doc\n\nGet the full content of a specific documentation page by its path.\n\n**Parameters:**\n- `path` (required): The file path to the documentation page (e.g., `app/getting-started/page.tsx`)\n\n**Example:**\n```json\n{\n  \"tool\": \"get_doc\",\n  \"params\": {\n    \"path\": \"app/deployment/page.tsx\"\n  }\n}\n```\n\n### list_docs\n\nList all available documentation pages, optionally filtered by directory.\n\n**Parameters:**\n- `directory` (optional): Optional directory to filter results (e.g., `components`)\n\n**Example:**\n```json\n{\n  \"tool\": \"list_docs\",\n  \"params\": {\n    \"directory\": \"configuration\"\n  }\n}\n```\n\n## How it works\n\nDoccupine's MCP server uses semantic search powered by vector embeddings to provide accurate, context-aware search results.\n\n### Indexing Process\n\n1. **Document Discovery**: The server scans your `app/` directory for all `page.tsx`, `page.ts`, `page.jsx`, and `page.js` files.\n2. **Content Extraction**: It extracts content from `const content =` declarations in your page files.\n3. **Chunking**: Large documents are split into chunks of approximately 800 characters with 100 character overlap for better context preservation.\n4. **Embedding Generation**: Each chunk is converted to a vector embedding using your configured LLM provider's embedding model.\n5. **Index Building**: All embeddings are stored in memory for fast similarity search.\n\n### Search Process\n\n1. **Query Embedding**: The search query is converted to a vector embedding using the same embedding model.\n2. **Similarity Calculation**: Cosine similarity is calculated between the query embedding and all document chunk embeddings.\n3. **Ranking**: Results are sorted by similarity score (highest first).\n4. **Response**: The top N results (based on the limit parameter) are returned with their paths, URIs, scores, and text content.\n\n<Callout type=\"note\">\n  The index is built automatically when the server starts. It is stored in memory and persists for the lifetime of the server process. If you update your documentation, restart the server to rebuild the index.\n</Callout>\n\n## Use your MCP server\n\nYour users must connect your MCP server to their preferred AI tools.\n\n1. Make your MCP server URL publicly available.\n2. Users copy your MCP server URL and add it to their tools.\n3. Users access your documentation through their tools.\n\nThese are some of the ways you can help your users connect to your MCP server:\n\n<Tabs>\n  <TabContent title=\"Claude\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the steps to connect it to Claude.\n        1. Navigate to the [Connectors](https://claude.ai/settings/connectors) page in the Claude settings.\n        2. Select **Add custom connector**.\n        3. Add your MCP server name and URL.\n        4. Select **Add**.\n        5. When using Claude, select the attachments button (the plus icon).\n        6. Select your MCP server.\n      </Step>\n    </Steps>\n    See the [Model Context Protocol documentation](https://modelcontextprotocol.io/docs/tutorials/use-remote-mcp-server#connecting-to-a-remote-mcp-server) for more details.\n  </TabContent>\n  <TabContent title=\"Claude Code\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the command to connect it to Claude Code.\n        ```bash\n        claude mcp add --transport http <name> <url>\n        ```\n      </Step>\n    </Steps>\n    See the [Claude Code documentation](https://docs.anthropic.com/en/docs/claude-code/mcp#installing-mcp-servers) for more details.\n  </TabContent>\n  <TabContent title=\"Cursor\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the steps to connect it to Cursor.\n        1. Use <kbd>Command</kbd> + <kbd>Shift</kbd> + <kbd>P</kbd> (<kbd>Ctrl</kbd> + <kbd>Shift</kbd> + <kbd>P</kbd> on Windows) to open the command palette.\n        2. Search for \"Open MCP settings\".\n        3. Select **Add custom MCP**. This opens the `mcp.json` file.\n        4. In `mcp.json`, configure your server:\n        ```json\n        {\n          \"mcpServers\": {\n            \"<your-mcp-server-name>\": {\n              \"url\": \"<your-mcp-server-url>\"\n            }\n          }\n        }\n        ```\n      </Step>\n    </Steps>\n    See the [Cursor documentation](https://docs.cursor.com/en/context/mcp#installing-mcp-servers) for more details.\n  </TabContent>\n  <TabContent title=\"VS Code\">\n    <Steps>\n      <Step title=\"Get your MCP server URL.\">\n        Your MCP server URL is available at `https://your-domain.com/api/mcp`.\n      </Step>\n      <Step title=\"Publish your MCP server URL for your users.\">\n        Create a guide for your users that includes your MCP server URL and the steps to connect it to VS Code.\n        1. Create a `.vscode/mcp.json` file.\n        2. In `mcp.json`, configure your server:\n        ```json\n        {\n          \"servers\": {\n            \"<your-mcp-server-name>\": {\n              \"type\": \"http\",\n              \"url\": \"<your-mcp-server-url>\"\n            }\n          }\n        }\n        ```\n      </Step>\n    </Steps>\n    See the [VS Code documentation](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) for more details.\n  </TabContent>\n</Tabs>\n\n## Requirements\n\nTo use the MCP server, you need to have the AI Assistant configured. The MCP server uses the same LLM configuration for generating embeddings.\n\n| Variable | Required | Description |\n|---|---|---|\n| `LLM_PROVIDER` | Yes | Your LLM provider (`openai`, `anthropic`, or `google`) |\n| `DOCS_API_KEY` | No | When set, requires Bearer token authentication on `/api/mcp` |\n\n<Callout type=\"warning\">\n  The MCP server requires an LLM provider to be configured for generating embeddings. Make sure you have set up your AI Assistant with a valid API key before using the MCP server.\n</Callout>\n\nSee the [AI Assistant documentation](/ai-assistant) for configuration details.\n\n## Content indexing\n\nYour MCP server searches content extracted from your page files. The server automatically discovers and indexes all `page.tsx`, `page.ts`, `page.jsx`, and `page.js` files in your `app/` directory.\n\n### Content extraction\n\nThe server extracts content from `const content =` declarations in your page files. Make sure your documentation pages export a `content` constant with your markdown or MDX content.\n\n**Example:**\n```tsx\nexport const content = `\n# Getting Started\n\nWelcome to the documentation...\n`;\n```\n\n### Excluded directories\n\nThe following directories are automatically excluded from indexing:\n- `node_modules`\n- `.next`\n- `.git`\n- `api`\n\n## Troubleshooting\n\n### Index not building\n\nIf the index is not building, check:\n- Your LLM provider is configured correctly in your `.env` file\n- You have a valid API key set\n- Your documentation pages export a `content` constant\n\n### No search results\n\nIf searches return no results:\n- Verify that your documentation pages are in the `app/` directory\n- Check that your pages export a `content` constant\n- Ensure the index has been built (check the `index.ready` status via GET `/api/mcp`)\n\n### Slow search performance\n\nThe first search may be slower as it builds the index. Subsequent searches are fast as they use the in-memory index. If performance is consistently slow:\n- Check your embedding API response times\n- Consider reducing the number of documentation pages\n- Verify your server has sufficient memory\n\n### Cloudflare blocking MCP requests\n\nIf you use Cloudflare as a proxy in front of your documentation site, Cloudflare's bot protection may block server-to-server requests from AI tools like Claude.ai. This can cause MCP connections to fail silently or return errors.\n\nThere are two ways to resolve this:\n\n**Option 1: Disable the Cloudflare proxy (simplest)**\n\nIn your Cloudflare DNS settings, click the orange cloud icon next to your domain record to switch it to \"DNS only\" (grey cloud). This disables Cloudflare's proxy and bot protection for your domain, allowing MCP requests to reach your server directly.\n\n**Option 2: Add a Cloudflare WAF exception (keeps your custom domain proxied)**\n\nIn Cloudflare dashboard:\n\n1. Go to **Security > WAF**.\n2. Click **Create rule**.\n3. Set it up as:\n   - **Rule name**: Allow MCP API\n   - **Field**: URI Path\n   - **Operator**: starts with\n   - **Value**: `/api/mcp`\n   - **Action**: Skip -- then check all remaining custom rules, Rate limiting rules, and Bot Fight Mode / Super Bot Fight Mode.\n4. Deploy the rule and make sure it is ordered first (above other rules).\n\n<Callout type=\"warning\">\n  Also check **Security > Bots** in your Cloudflare dashboard. If \"Bot Fight Mode\" or \"Super Bot Fight Mode\" is enabled, that is very likely what is blocking server-to-server requests from AI tools.\n</Callout>\n\n## Best practices\n\n* **Keep content up-to-date**: Restart your server after updating documentation to rebuild the index with fresh content.\n* **Use descriptive queries**: More specific queries yield better semantic search results.\n* **Monitor index status**: Use the GET endpoint to check if your index is ready before performing searches.\n* **Optimize content structure**: Well-structured markdown with clear headings improves search relevance.";

package/dist/templates/mdx/model-context-protocol.mdx.js

@@ -183,7 +183,7 @@ Doccupine's MCP server uses semantic search powered by vector embeddings to prov
 
 1. **Document Discovery**: The server scans your \`app/\` directory for all \`page.tsx\`, \`page.ts\`, \`page.jsx\`, and \`page.js\` files.
 2. **Content Extraction**: It extracts content from \`const content =\` declarations in your page files.
-3. **Chunking**: Large documents are split into chunks of approximately 1500 characters with 200 character overlap for better context preservation.
+3. **Chunking**: Large documents are split into chunks of approximately 800 characters with 100 character overlap for better context preservation.
 4. **Embedding Generation**: Each chunk is converted to a vector embedding using your configured LLM provider's embedding model.
 5. **Index Building**: All embeddings are stored in memory for fast similarity search.
 
@@ -195,7 +195,7 @@ Doccupine's MCP server uses semantic search powered by vector embeddings to prov
 4. **Response**: The top N results (based on the limit parameter) are returned with their paths, URIs, scores, and text content.
 
 <Callout type="note">
-  The index is built automatically when the first search is performed. The index is stored in memory and persists for the lifetime of the server process. If you update your documentation, restart the server to rebuild the index.
+  The index is built automatically when the server starts. It is stored in memory and persists for the lifetime of the server process. If you update your documentation, restart the server to rebuild the index.
 </Callout>
 
 ## Use your MCP server

package/dist/templates/mdx/navigation.mdx.d.ts

@@ -1 +1 @@
-export declare const navigationMdxTemplate = "---\ntitle: \"Navigation\"\ndescription: \"Organize and structure your navigation.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 2\n---\n# Navigation\nDoccupine builds your sidebar automatically from your MDX pages. By default, it reads the page frontmatter and groups pages into categories in the order you define. For larger docs, you can take full control with a `navigation.json` file.\n\n## Automatic navigation (default)\nWhen no custom navigation is provided, Doccupine generates a structure based on each page's frontmatter.\n\n### Frontmatter fields\n- **category**: The category name that groups the page in the sidebar.\n- **categoryOrder**: The position of the category within the sidebar. Lower numbers appear first.\n- **order**: The position of the page within its category. Lower numbers appear first.\n\n### Example frontmatter\n\n```text\n---\ntitle: \"Navigation\"\ndescription: \"Organize and structure your navigation.\"\ndate: \"2025-01-15\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 2\n---\n```\n\nThis approach is great for small sets of documents. For larger projects, setting these fields on every page can become repetitive.\n\n## Custom navigation with navigation.json\nTo centrally define the entire sidebar, create a `navigation.json` at your project root (the same folder where you execute `npx doccupine`). When present, it takes priority over page frontmatter and fully controls the navigation structure.\n\n### Array format\nThe simplest format is an array of categories. When using [sections](/sections), this applies to the root section only.\n\n```json\n[\n  {\n    \"label\": \"General\",\n    \"links\": [\n      { \"slug\": \"\", \"title\": \"Getting Started\" },\n      { \"slug\": \"commands\", \"title\": \"Commands\" }\n    ]\n  },\n  {\n    \"label\": \"Components\",\n    \"links\": [\n      { \"slug\": \"headers-and-text\", \"title\": \"Headers and Text\" },\n      { \"slug\": \"lists-and-tables\", \"title\": \"Lists and tables\" },\n      { \"slug\": \"code\", \"title\": \"Code\" },\n      { \"slug\": \"accordion\", \"title\": \"Accordion\" },\n      { \"slug\": \"tabs\", \"title\": \"Tabs\" },\n      { \"slug\": \"cards\", \"title\": \"Cards\" },\n      { \"slug\": \"buttons\", \"title\": \"Buttons\" },\n      { \"slug\": \"callouts\", \"title\": \"Callouts\" },\n      { \"slug\": \"images-and-embeds\", \"title\": \"Images and embeds\" },\n      { \"slug\": \"icons\", \"title\": \"Icons\" },\n      { \"slug\": \"fields\", \"title\": \"Fields\" },\n      { \"slug\": \"update\", \"title\": \"Update\" },\n      { \"slug\": \"columns\", \"title\": \"Columns\" },\n      { \"slug\": \"steps\", \"title\": \"Steps\" }\n    ]\n  },\n  {\n    \"label\": \"Configuration\",\n    \"links\": [\n      { \"slug\": \"globals\", \"title\": \"Globals\" },\n      { \"slug\": \"navigation\", \"title\": \"Navigation\" },\n      { \"slug\": \"sections\", \"title\": \"Sections\" },\n      { \"slug\": \"footer-links\", \"title\": \"Footer Links\" },\n      { \"slug\": \"theme\", \"title\": \"Theme\" },\n      { \"slug\": \"media-and-assets\", \"title\": \"Media and assets\" },\n      { \"slug\": \"fonts\", \"title\": \"Fonts\" },\n      { \"slug\": \"ai-assistant\", \"title\": \"AI Assistant\" },\n      { \"slug\": \"model-context-protocol\", \"title\": \"Model Context Protocol\" },\n      { \"slug\": \"deployment\", \"title\": \"Deployment\" }\n    ]\n  }\n]\n```\n\n### Object format (per-section)\nWhen using [sections](/sections), you can define navigation for each section by using an object keyed by section slug. Sections without a key fall back to auto-generated navigation from frontmatter.\n\n```json\n{\n  \"\": [\n    {\n      \"label\": \"General\",\n      \"links\": [\n        { \"slug\": \"\", \"title\": \"Getting Started\" },\n        { \"slug\": \"commands\", \"title\": \"Commands\" }\n      ]\n    }\n  ],\n  \"platform\": [\n    {\n      \"label\": \"Overview\",\n      \"links\": [\n        { \"slug\": \"platform/auth\", \"title\": \"Authentication\" },\n        { \"slug\": \"platform/users\", \"title\": \"Users\" }\n      ]\n    }\n  ]\n}\n```\n\nThe key `\"\"` controls the root section. Other keys match section slugs defined in `sections.json` or derived from frontmatter. See [Sections](/sections) for details on configuring sections.\n\n### Fields\n- **label**: The section header shown in the sidebar.\n- **links**: An array of page entries for that section.\n  - **slug**: The MDX file slug (filename without extension). Use an empty string `\"\"` for `index.mdx`.\n  - **title**: The display title in the navigation. This can differ from the page's `title` frontmatter.\n\n## Precedence and behavior\n\n<Callout type=\"note\">\n  `navigation.json` takes priority over frontmatter. If present, it fully controls the sidebar structure for the sections it covers.\n</Callout>\n\n- Without `navigation.json`, the sidebar is built from page frontmatter: `category` -> grouped; `categoryOrder` -> category position; `order` -> page position.\n- When using the object format, sections not listed in `navigation.json` fall back to frontmatter-based navigation.\n- Pages without a `category` appear at the top level.\n\n## Tips\n- **Start simple**: Use frontmatter for small docs. Switch to `navigation.json` as the structure grows.\n- **Keep slugs consistent**: `slug` must match the MDX filename (e.g., `text.mdx` -> `text`).\n- **Control titles**: Use `title` in `navigation.json` to customize sidebar labels without changing page frontmatter.\n- **Per-section navigation**: Use the object format to define different sidebars for each section. Mix and match - define some sections explicitly and let others auto-generate.";
+export declare const navigationMdxTemplate = "---\ntitle: \"Navigation\"\ndescription: \"Organize and structure your navigation.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 2\n---\n# Navigation\nDoccupine builds your sidebar automatically from your MDX pages. By default, it reads the page frontmatter and groups pages into categories in the order you define. For larger docs, you can take full control with a `navigation.json` file.\n\n## Automatic navigation (default)\nWhen no custom navigation is provided, Doccupine generates a structure based on each page's frontmatter.\n\n### Frontmatter fields\n- **category**: The category name that groups the page in the sidebar.\n- **categoryOrder**: The position of the category within the sidebar. Lower numbers appear first.\n- **order**: The position of the page within its category. Lower numbers appear first.\n\n### Example frontmatter\n\n```text\n---\ntitle: \"Navigation\"\ndescription: \"Organize and structure your navigation.\"\ndate: \"2025-01-15\"\ncategory: \"Configuration\"\ncategoryOrder: 3\norder: 2\n---\n```\n\nThis approach is great for small sets of documents. For larger projects, setting these fields on every page can become repetitive.\n\n## Custom navigation with navigation.json\nTo centrally define the entire sidebar, create a `navigation.json` at your project root (the same folder where you execute `npx doccupine`). When present, it takes priority over page frontmatter and fully controls the navigation structure.\n\n### Array format\nThe simplest format is an array of categories. When using [sections](/sections), this applies to the root section only.\n\n```json\n[\n  {\n    \"label\": \"General\",\n    \"links\": [\n      { \"slug\": \"\", \"title\": \"Getting Started\" },\n      { \"slug\": \"commands\", \"title\": \"Commands\" }\n    ]\n  },\n  {\n    \"label\": \"Components\",\n    \"links\": [\n      { \"slug\": \"headers-and-text\", \"title\": \"Headers and Text\" },\n      { \"slug\": \"lists-and-tables\", \"title\": \"Lists and tables\" },\n      { \"slug\": \"code\", \"title\": \"Code\" },\n      { \"slug\": \"accordion\", \"title\": \"Accordion\" },\n      { \"slug\": \"tabs\", \"title\": \"Tabs\" },\n      { \"slug\": \"cards\", \"title\": \"Cards\" },\n      { \"slug\": \"buttons\", \"title\": \"Buttons\" },\n      { \"slug\": \"callouts\", \"title\": \"Callouts\" },\n      { \"slug\": \"image-and-embeds\", \"title\": \"Images and embeds\" },\n      { \"slug\": \"icons\", \"title\": \"Icons\" },\n      { \"slug\": \"fields\", \"title\": \"Fields\" },\n      { \"slug\": \"update\", \"title\": \"Update\" },\n      { \"slug\": \"columns\", \"title\": \"Columns\" },\n      { \"slug\": \"steps\", \"title\": \"Steps\" }\n    ]\n  },\n  {\n    \"label\": \"Configuration\",\n    \"links\": [\n      { \"slug\": \"globals\", \"title\": \"Globals\" },\n      { \"slug\": \"navigation\", \"title\": \"Navigation\" },\n      { \"slug\": \"sections\", \"title\": \"Sections\" },\n      { \"slug\": \"footer-links\", \"title\": \"Footer Links\" },\n      { \"slug\": \"theme\", \"title\": \"Theme\" },\n      { \"slug\": \"media-and-assets\", \"title\": \"Media and assets\" },\n      { \"slug\": \"fonts\", \"title\": \"Fonts\" },\n      { \"slug\": \"ai-assistant\", \"title\": \"AI Assistant\" },\n      { \"slug\": \"model-context-protocol\", \"title\": \"Model Context Protocol\" },\n      { \"slug\": \"deployment\", \"title\": \"Deployment\" }\n    ]\n  }\n]\n```\n\n### Object format (per-section)\nWhen using [sections](/sections), you can define navigation for each section by using an object keyed by section slug. Sections without a key fall back to auto-generated navigation from frontmatter.\n\n```json\n{\n  \"\": [\n    {\n      \"label\": \"General\",\n      \"links\": [\n        { \"slug\": \"\", \"title\": \"Getting Started\" },\n        { \"slug\": \"commands\", \"title\": \"Commands\" }\n      ]\n    }\n  ],\n  \"platform\": [\n    {\n      \"label\": \"Overview\",\n      \"links\": [\n        { \"slug\": \"platform/auth\", \"title\": \"Authentication\" },\n        { \"slug\": \"platform/users\", \"title\": \"Users\" }\n      ]\n    }\n  ]\n}\n```\n\nThe key `\"\"` controls the root section. Other keys match section slugs defined in `sections.json` or derived from frontmatter. See [Sections](/sections) for details on configuring sections.\n\n### Fields\n- **label**: The section header shown in the sidebar.\n- **links**: An array of page entries for that section.\n  - **slug**: The MDX file slug (filename without extension). Use an empty string `\"\"` for `index.mdx`.\n  - **title**: The display title in the navigation. This can differ from the page's `title` frontmatter.\n\n## Precedence and behavior\n\n<Callout type=\"note\">\n  `navigation.json` takes priority over frontmatter. If present, it fully controls the sidebar structure for the sections it covers.\n</Callout>\n\n- Without `navigation.json`, the sidebar is built from page frontmatter: `category` -> grouped; `categoryOrder` -> category position; `order` -> page position.\n- When using the object format, sections not listed in `navigation.json` fall back to frontmatter-based navigation.\n- Pages without a `category` appear at the top level.\n\n## Tips\n- **Start simple**: Use frontmatter for small docs. Switch to `navigation.json` as the structure grows.\n- **Keep slugs consistent**: `slug` must match the MDX filename (e.g., `text.mdx` -> `text`).\n- **Control titles**: Use `title` in `navigation.json` to customize sidebar labels without changing page frontmatter.\n- **Per-section navigation**: Use the object format to define different sidebars for each section. Mix and match - define some sections explicitly and let others auto-generate.";

package/dist/templates/mdx/navigation.mdx.js

@@ -58,7 +58,7 @@ The simplest format is an array of categories. When using [sections](/sections),
       { "slug": "cards", "title": "Cards" },
       { "slug": "buttons", "title": "Buttons" },
       { "slug": "callouts", "title": "Callouts" },
-      { "slug": "images-and-embeds", "title": "Images and embeds" },
+      { "slug": "image-and-embeds", "title": "Images and embeds" },
       { "slug": "icons", "title": "Icons" },
       { "slug": "fields", "title": "Fields" },
       { "slug": "update", "title": "Update" },

package/dist/templates/mdx/platform/ai-assistant.mdx.d.ts

@@ -1 +1 @@
-export declare const platformAiAssistantMdxTemplate = "---\ntitle: \"AI Assistant\"\ndescription: \"Configure the built-in AI assistant that ships with every Doccupine documentation site.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 2\norder: 6\nsection: \"Platform\"\n---\n# AI Assistant\nEvery Doccupine site ships with a built-in AI assistant that helps visitors find answers across your documentation. The AI settings page lets you choose how it's powered.\n\n## Modes\n\n### Platform (default)\nUses Doccupine's built-in integration. Zero configuration needed - the AI assistant works out of the box with no API keys or setup.\n\n### Custom\nBring your own API key for full control over the AI model. Supported providers:\n\n- **OpenAI**\n- **Anthropic**\n- **Google**\n\nIn Custom mode, you can also configure:\n- **Embedding model** - the model used to index your documentation content\n- **Temperature** - controls response creativity (0.0 for focused answers, up to 1.0 for more varied responses)\n\nFor a complete list of available models, refer to the official documentation of your chosen provider.\n\n### Off\nCompletely disables the AI assistant on your site.\n\n<Callout type=\"warning\">\n  AI settings are stored as environment variables on your deployment, not in a JSON file. After saving, a redeploy is triggered automatically to apply the changes.\n</Callout>\n\n## MCP server authentication\nEvery Doccupine site exposes an MCP (Model Context Protocol) endpoint at `/api/mcp`. This lets external AI tools query your documentation programmatically.\n\nYou can set an optional **API key** to restrict access to the MCP endpoint. When set, requests must include the key in their authorization header.\n\n<Callout type=\"note\">\n  For more details on how the MCP endpoint works and how to connect it to AI tools, see the [Model Context Protocol documentation](/model-context-protocol).\n</Callout>";
+export declare const platformAiAssistantMdxTemplate = "---\ntitle: \"AI Assistant\"\ndescription: \"Configure the built-in AI assistant that ships with every Doccupine documentation site.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 2\norder: 6\nsection: \"Platform\"\n---\n# AI Assistant\nEvery Doccupine site ships with a built-in AI assistant that helps visitors find answers across your documentation. The AI settings page lets you choose how it's powered.\n\n## Modes\n\n### Platform (default)\nUses Doccupine's built-in integration. Zero configuration needed - the AI assistant works out of the box with no API keys or setup.\n\nEach plan includes a monthly AI usage budget:\n\n| Plan       | Monthly Budget |\n| ---------- | -------------- |\n| Trial      | $2             |\n| Pro        | $20            |\n| Enterprise | $50            |\n\nThe AI settings page shows a usage dashboard with your current spending and remaining budget. Usage resets automatically with your billing cycle.\n\n#### AI credit top-ups\n\nIf you run out of AI credits before your billing cycle resets, you can purchase a one-time top-up to increase your monthly limit. Available tiers:\n\n- **$5**\n- **$10**\n- **$20**\n\nTop-ups are added to your current cycle's budget immediately after purchase and reset when your billing cycle renews. You can purchase multiple top-ups in the same cycle.\n\n### Custom\nBring your own API key for full control over the AI model. Supported providers:\n\n- **OpenAI**\n- **Anthropic**\n- **Google**\n\nIn Custom mode, you can also configure:\n- **Embedding model** - the model used to index your documentation content\n- **Temperature** - controls response creativity (0.0 for focused answers, up to 1.0 for more varied responses)\n\nFor a complete list of available models, refer to the official documentation of your chosen provider.\n\n### Off\nCompletely disables the AI assistant on your site.\n\n<Callout type=\"warning\">\n  AI settings are stored as environment variables on your deployment, not in a JSON file. After saving, a redeploy is triggered automatically to apply the changes.\n</Callout>\n\n## MCP server authentication\nEvery Doccupine site exposes an MCP (Model Context Protocol) endpoint at `/api/mcp`. This lets external AI tools query your documentation programmatically.\n\nYou can set an optional **API key** to restrict access to the MCP endpoint. When set, requests must include the key in their authorization header.\n\n<Callout type=\"note\">\n  For more details on how the MCP endpoint works and how to connect it to AI tools, see the [Model Context Protocol documentation](/model-context-protocol).\n</Callout>";

package/dist/templates/mdx/platform/ai-assistant.mdx.js

@@ -15,6 +15,26 @@ Every Doccupine site ships with a built-in AI assistant that helps visitors find
 ### Platform (default)
 Uses Doccupine's built-in integration. Zero configuration needed - the AI assistant works out of the box with no API keys or setup.
 
+Each plan includes a monthly AI usage budget:
+
+| Plan       | Monthly Budget |
+| ---------- | -------------- |
+| Trial      | $2             |
+| Pro        | $20            |
+| Enterprise | $50            |
+
+The AI settings page shows a usage dashboard with your current spending and remaining budget. Usage resets automatically with your billing cycle.
+
+#### AI credit top-ups
+
+If you run out of AI credits before your billing cycle resets, you can purchase a one-time top-up to increase your monthly limit. Available tiers:
+
+- **$5**
+- **$10**
+- **$20**
+
+Top-ups are added to your current cycle's budget immediately after purchase and reset when your billing cycle renews. You can purchase multiple top-ups in the same cycle.
+
 ### Custom
 Bring your own API key for full control over the AI model. Supported providers:
 

package/dist/templates/mdx/platform/external-links.mdx.d.ts

@@ -1 +1 @@
-export declare const platformExternalLinksMdxTemplate = "---\ntitle: \"External Links\"\ndescription: \"Add quick-access link buttons to your site's footer for GitHub, Discord, and other external resources.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 2\norder: 5\nsection: \"Platform\"\n---\n# External Links\nThe Footer Links settings page lets you add external link buttons to your documentation site's footer. These provide quick access to your project's GitHub repository, Discord server, social profiles, and other resources.\n\n## Adding a link\nClick **Add Link** and configure:\n\n- **Title** - the display text for the link\n- **URL** - the target URL\n- **Icon** (optional) - choose from one of the preset icons\n\n## Available icons\nYou can assign one of these icons to each link:\n\n- GitHub\n- Twitter / X\n- Discord\n- Slack\n- LinkedIn\n- YouTube\n- Website (globe)\n- Email\n- Docs (book)\n- Code\n- Package\n- RSS\n- Chat\n- Sponsor (heart)\n\nLeave the icon unset for a text-only link.\n\n## How it works\nLink settings are stored in `links.json` at the root of your repository:\n\n```json\n[\n  {\n    \"title\": \"GitHub\",\n    \"url\": \"https://github.com/your-org/your-repo\",\n    \"icon\": \"github\"\n  },\n  {\n    \"title\": \"Discord\",\n    \"url\": \"https://discord.gg/your-invite\",\n    \"icon\": \"discord\"\n  }\n]\n```";
+export declare const platformExternalLinksMdxTemplate = "---\ntitle: \"External Links\"\ndescription: \"Add quick-access link buttons to your site's footer for GitHub, Discord, and other external resources.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 2\norder: 5\nsection: \"Platform\"\n---\n# External Links\nThe Footer Links settings page lets you add external link buttons to your documentation site's footer. These provide quick access to your project's GitHub repository, Discord server, social profiles, and other resources.\n\n## Adding a link\nClick **Add Link** and configure:\n\n- **Title** - the display text for the link\n- **URL** - the target URL\n- **Icon** (optional) - choose from one of the preset icons\n\n## Available icons\nYou can assign one of these icons to each link:\n\n- GitHub\n- Twitter / X\n- Discord\n- Slack\n- LinkedIn\n- YouTube\n- Website (globe)\n- Email\n- Docs (book)\n- Code\n- Package\n- RSS\n- Chat\n- Sponsor (heart)\n\nLeave the icon unset for a text-only link.\n\nIf you edit `links.json` directly, use [Lucide](https://lucide.dev/) icon names (e.g. `github`, `message-circle`, `heart`).\n\n## How it works\nLink settings are stored in `links.json` at the root of your repository:\n\n```json\n[\n  {\n    \"title\": \"GitHub\",\n    \"url\": \"https://github.com/your-org/your-repo\",\n    \"icon\": \"github\"\n  },\n  {\n    \"title\": \"Discord\",\n    \"url\": \"https://discord.gg/your-invite\",\n    \"icon\": \"discord\"\n  }\n]\n```";

package/dist/templates/mdx/platform/external-links.mdx.js

@@ -37,6 +37,8 @@ You can assign one of these icons to each link:
 
 Leave the icon unset for a text-only link.
 
+If you edit \`links.json\` directly, use [Lucide](https://lucide.dev/) icon names (e.g. \`github\`, \`message-circle\`, \`heart\`).
+
 ## How it works
 Link settings are stored in \`links.json\` at the root of your repository:
 

package/dist/templates/mdx/platform/fonts-settings.mdx.d.ts

@@ -1 +1 @@
-export declare const platformFontsSettingsMdxTemplate = "---\ntitle: \"Fonts Settings\"\ndescription: \"Configure custom typography with Google Fonts or local font files.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 2\norder: 4\nsection: \"Platform\"\n---\n# Fonts Settings\nThe Fonts settings page lets you customize your documentation site's typography using Google Fonts or locally uploaded font files.\n\n## Google Fonts\nSelect a font from the full Google Fonts library:\n\n1. Type a font name to search the library.\n2. Select the **weights** you need (e.g., 400 for regular, 700 for bold). You can select multiple weights.\n3. Choose **subsets** for language support (latin, cyrillic, greek, vietnamese, etc.).\n\n<Callout type=\"note\">\n  Only include the weights and subsets you actually use. Each addition increases page load time.\n</Callout>\n\n## Local fonts\nUpload your own font files for complete typographic control:\n\n1. Click **Add Font Source** to add a font file entry.\n2. Upload a font file (WOFF2, WOFF, TTF, OTF, or EOT format).\n3. Set the **weight** (e.g., 400, 700) and **style** (normal or italic).\n4. Add more sources for additional weights and styles.\n\nWOFF2 is recommended for the best compression and browser support.\n\n## How it works\nFont settings are stored in `fonts.json` at the root of your repository. Here's an example using Google Fonts:\n\n```json\n{\n  \"type\": \"google\",\n  \"name\": \"Inter\",\n  \"weights\": [400, 500, 600, 700],\n  \"subsets\": [\"latin\"]\n}\n```";
+export declare const platformFontsSettingsMdxTemplate = "---\ntitle: \"Fonts Settings\"\ndescription: \"Configure custom typography with Google Fonts or local font files.\"\ndate: \"2026-02-19\"\ncategory: \"Configuration\"\ncategoryOrder: 2\norder: 4\nsection: \"Platform\"\n---\n# Fonts Settings\nThe Fonts settings page lets you customize your documentation site's typography using Google Fonts or locally uploaded font files.\n\n## Google Fonts\nSelect a font from the full Google Fonts library:\n\n1. Type a font name to search the library.\n2. Select the **weights** you need (e.g., 400 for regular, 700 for bold). You can select multiple weights.\n3. Choose **subsets** for language support (latin, cyrillic, greek, vietnamese, etc.).\n\n<Callout type=\"note\">\n  Only include the weights and subsets you actually use. Each addition increases page load time.\n</Callout>\n\n## Local fonts\nUpload your own font files for complete typographic control:\n\n1. Click **Add Font Source** to add a font file entry.\n2. Upload a font file (WOFF2, WOFF, TTF, OTF, or EOT format).\n3. Set the **weight** (e.g., 400, 700) and **style** (normal or italic).\n4. Add more sources for additional weights and styles.\n\nWOFF2 is recommended for the best compression and browser support.\n\n## How it works\nFont settings are stored in `fonts.json` at the root of your repository. Here's an example using Google Fonts:\n\n```json\n{\n  \"googleFont\": {\n    \"fontName\": \"Inter\",\n    \"subsets\": [\"latin\"],\n    \"weight\": [\"400\", \"500\", \"600\", \"700\"]\n  }\n}\n```\n\nSee the [Fonts](/fonts) page for the full configuration format, including local font support.";