umbraco
diff --git a/‎README.md‎
Lines changed: 63 additions & 0 deletions b/‎README.md‎
Lines changed: 63 additions & 0 deletions
diff --git a/‎docs/proto-docs/universal-media-upload.md‎
Lines changed: 125 additions & 0 deletions b/‎docs/proto-docs/universal-media-upload.md‎
Lines changed: 125 additions & 0 deletions
diff --git a/‎docs/tool-collection-filtering.md‎
Lines changed: 7 additions & 7 deletions b/‎docs/tool-collection-filtering.md‎
Lines changed: 7 additions & 7 deletions
diff --git a/‎jest.setup.ts‎
Lines changed: 10 additions & 2 deletions b/‎jest.setup.ts‎
Lines changed: 10 additions & 2 deletions
@@ -80,6 +80,34 @@ claude mcp list
 
 This will add umbraco-mcp to the existing project in the claude.json config file.
 
+#### Configuration via .mcp.json (Project-specific)
+
+For project-specific Claude Code configuration, create a `.mcp.json` file in your project root that references environment variables for sensitive data:
+
+```json
+{
+  "mcpServers": {
+    "umbraco-mcp": {
+      "command": "npx",
+      "args": ["@umbraco-cms/mcp-dev@beta"],
+      "env": {
+        "NODE_TLS_REJECT_UNAUTHORIZED": "0",
+        "UMBRACO_CLIENT_ID": "umbraco-back-office-mcp",
+        "UMBRACO_CLIENT_SECRET": "your-client-secret-here",
+        "UMBRACO_BASE_URL": "https://localhost:44391",
+        "UMBRACO_INCLUDE_TOOL_COLLECTIONS": "culture,document,media",
+        "UMBRACO_EXCLUDE_TOOLS": "delete-document,empty-recycle-bin"
+      }
+    }
+  }
+}
+```
+
+Using the `.mcp.json` file allows you to:
+- Configure MCP servers per project
+- Share configuration with team members (commit to version control)
+- Override global Claude Code MCP settings for specific projects
+- Move the environment varaibles to a .env file to prevent leaking of secrets to your code repo
 
 </details>
 
@@ -142,6 +170,7 @@ Add the following to the config file and update the env variables.
 ```
 </details>
 
+
 #### Authentication Configuration Keys
 
 - `UMBRACO_CLIENT_ID`
@@ -156,6 +185,40 @@ Umbraco API User client secert
 
 Url of the Umbraco site, it only needs to be the scheme and domain e.g https://<nolink/>example.com
 
+### Environment Configuration Options
+
+The Umbraco MCP server supports environment configuration via:
+1. **Environment variables in MCP client config as above** (Claude Desktop, VS Code, etc.)
+2. **Local `.env` file** for development (see `.env.example`)
+3. **CLI arguments** when running directly
+
+**Configuration precedence:** CLI arguments > Environment variables > `.env` file
+
+#### Using a `.env` file (Development)
+
+For local development, you can create a `.env` file in the project root:
+
+```bash
+# Edit with your values
+UMBRACO_CLIENT_ID=your-api-user-id
+UMBRACO_CLIENT_SECRET=your-api-secret
+UMBRACO_BASE_URL=http://localhost:56472
+```
+
+The `.env` file is gitignored to keep your secrets secure.
+
+#### CLI Arguments
+
+You can also pass configuration via CLI arguments:
+
+```bash
+npx @umbraco-cms/mcp-dev@beta \
+  --umbraco-client-id="your-id" \
+  --umbraco-client-secret="your-secret" \
+  --umbraco-base-url="http://localhost:56472" \
+  --env="/path/to/custom/.env"
+```
+
 ## API Coverage
 
 This MCP server provides **comprehensive coverage** of the Umbraco Management API. We have achieved **full parity** with all applicable endpoints, implementing tools for every operational endpoint suitable for AI-assisted content management.
 
@@ -0,0 +1,125 @@
+# Universal Media Upload Implementation
+
+## Overview
+
+The media upload system has been redesigned to provide a unified, simplified interface for uploading any type of media file to Umbraco. This simplifies the standard two-step process (create temporary file → create media) with a single tool call that handles all media types.
+
+## Architecture Decision: Single Universal Tool
+
+Instead of creating separate tools for each media type (create-image, create-pdf, create-video, etc.), we implemented a **single universal tool** that:
+- Accepts any media type via an explicit `mediaTypeName` parameter
+- Trusts the LLM to specify the correct media type based on context
+- Only validates SVG files (the one exception where file type matters for technical reasons)
+- Supports custom media types created in Umbraco via dynamic API lookup
+
+### Why Trust the LLM?
+
+**Advantages:**
+- ✅ LLMs understand semantic context better than file extensions
+- ✅ Simpler implementation - no complex extension mapping tables
+- ✅ Works seamlessly with custom media types
+- ✅ Explicit is better than implicit
+- ✅ Dynamic lookup ensures compatibility with any Umbraco installation
+
+**The Only Exception - SVG:**
+- SVGs can be mistaken for images by LLMs
+- File extension check is simple and reliable for this one case
+- Auto-correct prevents technical errors (SVG uploaded as "Image" type fails in Umbraco)
+
+## Tools Implemented
+
+### 1. create-media
+
+**Purpose:** Upload any single media file to Umbraco
+
+**Schema:**
+```typescript
+{
+  sourceType: "filePath" | "url" | "base64",
+  name: string,
+  mediaTypeName: string,  // Required: explicit media type
+  filePath?: string,      // Required if sourceType = "filePath"
+  fileUrl?: string,       // Required if sourceType = "url"
+  fileAsBase64?: string,  // Required if sourceType = "base64"
+  parentId?: string       // Optional: parent folder UUID
+}
+```
+
+**Supported Media Types:**
+- **Image** - jpg, png, gif, webp (supports cropping features)
+- **Article** - pdf, docx, doc
+- **Audio** - mp3, wav, etc.
+- **Video** - mp4, webm, etc.
+- **Vector Graphic (SVG)** - svg files only
+- **File** - any other file type
+- **Custom** - any custom media type name created in Umbraco
+
+**Source Types:**
+1. **filePath** - Most efficient, zero token overhead, works with any size file
+2. **url** - Fetch from web URL
+3. **base64** - Only for small files (<10KB) due to token usage
+
+**Example Usage:**
+```typescript
+// Upload an image from local filesystem
+{
+  sourceType: "filePath",
+  name: "Product Photo",
+  mediaTypeName: "Image",
+  filePath: "/path/to/image.jpg"
+}
+
+// Upload a PDF from URL
+{
+  sourceType: "url",
+  name: "Annual Report",
+  mediaTypeName: "Article",
+  fileUrl: "https://example.com/report.pdf"
+}
+
+// Upload small image as base64
+{
+  sourceType: "base64",
+  name: "Icon",
+  mediaTypeName: "Image",
+  fileAsBase64: "iVBORw0KGgoAAAANS..."
+}
+```
+
+### 2. create-media-multiple
+
+**Purpose:** Batch upload multiple media files (maximum 20 per batch)
+
+**Schema:**
+```typescript
+{
+  sourceType: "filePath" | "url",  // No base64 for batch uploads
+  files: Array<{
+    name: string,
+    filePath?: string,
+    fileUrl?: string,
+    mediaTypeName?: string  // Optional per-file override, defaults to "File"
+  }>,
+  parentId?: string  // Optional: parent folder for all files
+}
+```
+
+**Features:**
+- Sequential processing to avoid API overload
+- Continue-on-error strategy - individual failures don't stop the batch
+- Returns detailed results per file with success/error status
+- Validates 20-file batch limit
+
+**Example Usage:**
+```typescript
+{
+  sourceType: "filePath",
+  files: [
+    { name: "Photo 1", filePath: "/path/to/photo1.jpg", mediaTypeName: "Image" },
+    { name: "Photo 2", filePath: "/path/to/photo2.jpg", mediaTypeName: "Image" },
+    { name: "Document", filePath: "/path/to/doc.pdf", mediaTypeName: "Article" }
+  ],
+  parentId: "parent-folder-id"
+}
+```
+
@@ -89,16 +89,16 @@ Some collections have dependencies that are automatically resolved:
 
 ### Configuration Loading
 
-Configuration is loaded from environment variables with automatic parsing:
+Configuration is loaded from the server config object:
 
 ```typescript
 export class CollectionConfigLoader {
-  static loadFromEnv(): CollectionConfiguration {
+  static loadFromConfig(config: UmbracoServerConfig): CollectionConfiguration {
     return {
-      enabledCollections: env.UMBRACO_INCLUDE_TOOL_COLLECTIONS ?? DEFAULT_COLLECTION_CONFIG.enabledCollections,
-      disabledCollections: env.UMBRACO_EXCLUDE_TOOL_COLLECTIONS ?? DEFAULT_COLLECTION_CONFIG.disabledCollections,
-      enabledTools: env.UMBRACO_INCLUDE_TOOLS ?? DEFAULT_COLLECTION_CONFIG.enabledTools,
-      disabledTools: env.UMBRACO_EXCLUDE_TOOLS ?? DEFAULT_COLLECTION_CONFIG.disabledTools,
+      enabledCollections: config.includeToolCollections ?? DEFAULT_COLLECTION_CONFIG.enabledCollections,
+      disabledCollections: config.excludeToolCollections ?? DEFAULT_COLLECTION_CONFIG.disabledCollections,
+      enabledTools: config.includeTools ?? DEFAULT_COLLECTION_CONFIG.enabledTools,
+      disabledTools: config.excludeTools ?? DEFAULT_COLLECTION_CONFIG.disabledTools,
     };
   }
 }
@@ -108,7 +108,7 @@ export class CollectionConfigLoader {
 
 The `UmbracoToolFactory` processes configuration and loads tools:
 
-1. Load configuration from environment variables
+1. Load configuration from server config
 2. Validate collection names and dependencies
 3. Resolve collection dependencies automatically
 4. Filter collections based on configuration
 
@@ -1,4 +1,12 @@
 import dotenv from 'dotenv';
+import { initializeUmbracoAxios } from './src/orval/client/umbraco-axios.js';
 
-// Load environment variables from .env.test
-dotenv.config({ path: '.env' });
+// Load environment variables from .env
+dotenv.config({ path: '.env' });
+
+// Initialize Umbraco Axios client with environment variables
+initializeUmbracoAxios({
+  clientId: process.env.UMBRACO_CLIENT_ID || '',
+  clientSecret: process.env.UMBRACO_CLIENT_SECRET || '',
+  baseUrl: process.env.UMBRACO_BASE_URL || ''
+});