@crownpeak/dqm-react-component-dev-mcp 1.2.4 → 1.3.0

package/data/CHANGELOG.md

@@ -5,6 +5,42 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.3.0] - 2026-01-09
+
+### Added
+- **GPT-5.2 Support**: Default AI model upgraded from `gpt-4.1-mini` to `gpt-5.2`
+  - 1M token context window enables larger translation batches and fewer API calls
+  - New `reasoning_effort` parameter (`low`/`medium`/`high`) for GPT-5 models
+  - Automatic API adaptation for GPT-5 vs GPT-4 differences:
+    - Token parameter: `max_completion_tokens` (GPT-5) vs `max_tokens` (GPT-4)
+    - Structured output: `json_schema` (GPT-5) vs `json_object` (GPT-4)
+    - System role: `developer` (GPT-5) vs `system` (GPT-4)
+  - 50,000 token context budget (vs 12,000 for GPT-4o)
+- **Model Capabilities Module**: New `src/utils/modelCapabilities.ts` for centralized model detection
+  - `isGPT5Model()`, `isReasoningModel()` utility functions
+  - `getModelCapabilities()` returns full capability info per model
+  - `buildTokenParams()`, `buildReasoningParams()`, `getSystemRole()` helpers
+- **Reasoning Effort UI**: New setting in AI Settings dialog (only visible for GPT-5 models)
+  - Persisted to `localStorage` as `dqm_reasoning_effort`
+  - Full i18n support (EN/DE/ES)
+- **MCP Server Feature**: Added MCP Server section to website features
+  - New feature card with link to documentation
+  - Translations for all 3 languages
+
+### Changed
+- **AI Settings Dialog**: GPT-5.2 now shown with "NEW" badge in model selector
+- **AI Summary Card**: Improved layout with column flex direction for better readability
+- **Website Hero Badge**: Updated from version number to "GPT-5.2 Powered AI"
+- **Website MSW Provider**: Refactored to `useMSW()` hook for better lifecycle control
+  - MSW now starts only when sidebar opens
+  - MSW resets completely when sidebar closes
+- **Footer Links**: Updated Crownpeak DQM Platform URL to German FirstSpirit product page
+
+### Fixed
+- **Translation Feature Descriptions**: Updated to reflect GPT-5.2 capabilities
+- **Website Feature Cards**: Added MCP Server feature with external documentation link
+- **Accessibility**: Added `aria-label` to MCP Server feature link
+
 ## [1.2.4] - 2026-01-08
 
 ### Added

package/data/docs/AI-FEATURES.md

@@ -69,10 +69,46 @@ flowchart TD
 ### OpenAI Backend
 
 **Models Supported**:
-- `gpt-4o-mini` (Recommended, fast & cheap)
-- `gpt-4o` (Higher quality)
-- `gpt-4.1-mini`
-- `gpt-4.1`
+
+| Model | Context Window | Recommended For |
+|-------|----------------|-----------------|
+| `gpt-5.2` | 1M tokens | 🌟 Default – Best performance & quality |
+| `gpt-4.1-mini` | 1M tokens | Cost-effective alternative |
+| `gpt-4.1` | 1M tokens | High quality GPT-4 |
+| `gpt-4o-mini` | 128K tokens | Fast & cheap (legacy) |
+| `gpt-4o` | 128K tokens | Higher quality (legacy) |
+
+> **🚀 GPT-5.2** is the new default model with 1M token context window, enabling much larger translation batches and better quality.
+
+#### GPT-5 vs GPT-4 Differences
+
+The component automatically adapts to API differences between GPT-5 and GPT-4 models:
+
+| Feature | GPT-5.x | GPT-4.x |
+|---------|---------|---------|
+| Token Parameter | `max_completion_tokens` | `max_tokens` |
+| Structured Output | `json_schema` | `json_object` |
+| System Role | `developer` | `system` |
+| Reasoning Effort | ✅ Supported | ❌ Not available |
+| Context Budget | 50,000 tokens | 12,000 tokens |
+
+#### Reasoning Effort (GPT-5 Only)
+
+GPT-5 models support a **Reasoning Effort** parameter that controls how deeply the model analyzes translations:
+
+| Level | Description | Best For |
+|-------|-------------|----------|
+| `low` | Fast, straightforward translations | Simple content, high volume |
+| `medium` | Balanced analysis (default) | Most use cases |
+| `high` | Deep reasoning, highest accuracy | Technical content, quality-critical |
+
+Configure via the **AI Settings** dialog or localStorage:
+
+```typescript
+localStorage.setItem('dqm_reasoning_effort', 'medium'); // 'low' | 'medium' | 'high'
+```
+
+> **Note:** Reasoning Effort is only visible in the UI when a GPT-5 model is selected.
 
 **Setup**:
 
@@ -284,6 +320,8 @@ graph TD
 - **Tiny**: 20+ checkpoints → Top 5 most critical only
 - **Fail**: > 50 checkpoints → Skip summary (too large)
 
+> **💡 GPT-5 Advantage**: With GPT-5.2's 1M token context window, the component uses a 50,000 token context budget (vs 12,000 for GPT-4o), enabling larger batches and fewer API calls.
+
 **Prompt Template**:
 
 ```typescript
@@ -397,7 +435,8 @@ The AI features store configuration in `localStorage`:
 | `dqm_translate_results_mode` | `'fast' \| 'full'` | `'fast'` | Translation mode |
 | `dqm_ai_summary_enabled` | `'true' \| 'false'` | `'true'` | Summary toggle |
 | `dqm_openai_apiKey` | `string` | `null` | OpenAI API key |
-| `dqm_openai_model` | `string` | `'gpt-4.1-mini'` | OpenAI model name |
+| `dqm_openai_model` | `string` | `'gpt-5.2'` | OpenAI model name |
+| `dqm_reasoning_effort` | `'low' \| 'medium' \| 'high'` | `'medium'` | GPT-5 reasoning depth |
 | `dqm_openai_baseUrl` | `string` | `'https://api.openai.com/v1'` | OpenAI base URL |
 
 **Access Pattern**:
@@ -439,6 +478,24 @@ setLocalStorageItem('dqm_openai_apiKey', 'sk-...');
 
 ### Cost Estimation (OpenAI)
 
+#### GPT-5.2 (Default)
+
+Based on `gpt-5.2` pricing (~$0.10/1M input tokens, ~$0.40/1M output tokens):
+
+| Operation | Average Tokens | Cost per Call |
+|-----------|----------------|---------------|
+| Translate 1 checkpoint | ~200 input, ~100 output | ~$0.00006 |
+| Translate 50 checkpoints | ~10k input, ~5k output | ~$0.003 |
+| Summary (5 issues) | ~500 input, ~200 output | ~$0.00013 |
+| Summary (20 issues) | ~2k input, ~500 output | ~$0.0004 |
+
+**Monthly costs** (assuming 1000 analyses/month with 20 checkpoints each):
+- Translation only: ~$3.00/month
+- Summary only: ~$0.40/month
+- Both: ~$3.40/month
+
+#### GPT-4o-mini (Legacy)
+
 Based on `gpt-4o-mini` pricing (~$0.15/1M input tokens, ~$0.60/1M output tokens):
 
 | Operation | Average Tokens | Cost per Call |

package/data/docs/API-REFERENCE.md

@@ -334,7 +334,7 @@ function MyComponent() {
     const engine = useAIEngine({
         enabled: true,
         openAiApiKey: 'sk-...',
-        openAiModel: 'gpt-4.1-mini',        // Optional, default: 'gpt-4.1-mini'
+        openAiModel: 'gpt-5.2',        // Optional, default: 'gpt-5.2'
         openAiBaseUrl: 'https://api.openai.com/v1',  // Optional
     });
 }
@@ -346,7 +346,7 @@ function MyComponent() {
 |----------|------|----------|-------------|
 | `enabled` | `boolean` | ✅ | Whether AI features are enabled |
 | `openAiApiKey` | `string` | ❌ | OpenAI API key |
-| `openAiModel` | `string` | ❌ | OpenAI model name (default: 'gpt-4.1-mini') |
+| `openAiModel` | `string` | ❌ | OpenAI model name (default: 'gpt-5.2') |
 | `openAiBaseUrl` | `string` | ❌ | OpenAI base URL (default: 'https://api.openai.com/v1') |
 
 ##### Returns (UseAIEngineReturn)
@@ -386,7 +386,7 @@ function MyComponent() {
         cacheManager,
         originalData: analysisData,
         targetLang: 'de',
-        modelId: 'gpt-4.1-mini',
+        modelId: 'gpt-5.2',
         enabled: true,
         mode: 'fast',
         computeBudgetMs: 15000,
@@ -466,7 +466,7 @@ function MyComponent() {
         engine,
         originalData: analysisData,
         targetLang: 'de',
-        modelId: 'gpt-4.1-mini',
+        modelId: 'gpt-5.2',
         enabled: true,
         cache: cacheManager.cache,
     });
@@ -988,7 +988,7 @@ The DQM component uses localStorage for persisting user preferences and authenti
 | Key | Type | Description |
 |-----|------|-------------|
 | `dqm_openai_apiKey` | `string` | OpenAI API key for translation/summary |
-| `dqm_openai_model` | `string` | OpenAI model (default: 'gpt-4.1-mini') |
+| `dqm_openai_model` | `string` | OpenAI model (default: 'gpt-5.2') |
 | `dqm_target_language` | `string` | Target language for translation (ISO 639-1) |
 | `dqm_translate_results_enabled` | `'true' \| 'false'` | Translation feature enabled |
 | `dqm_ai_summary_enabled` | `'true' \| 'false'` | AI summary feature enabled |
@@ -1007,7 +1007,7 @@ The DQM component uses localStorage for persisting user preferences and authenti
 ```typescript
 // Set OpenAI configuration before loading DQM
 localStorage.setItem('dqm_openai_apiKey', 'sk-...');
-localStorage.setItem('dqm_openai_model', 'gpt-4o-mini');
+localStorage.setItem('dqm_openai_model', 'gpt-5.2');
 localStorage.setItem('dqm_target_language', 'de');
 localStorage.setItem('dqm_translate_results_enabled', 'true');
 ```

package/data/docs/EXAMPLES.md

@@ -419,11 +419,11 @@ export default App;
 ```
 
 **Key Features:**
-- **Backend:** OpenAI (gpt-4o-mini, gpt-4o, gpt-4.1)
+- **Backend:** OpenAI (gpt-5.2, gpt-4o, gpt-4.1)
 - **Performance:** ~2-5s for 50 checkpoints (batch processing with 3 concurrent requests)
 - **Caching:** Automatic IndexedDB + In-Memory caching (FNV-1a hash-based)
 - **Mode:** `fast` (15s timeout, fails gracefully) or `full` (120s timeout, complete translation)
-- **Cost:** ~$0.001-0.003 per checkpoint (gpt-4o-mini)
+- **Cost:** ~$0.001-0.003 per checkpoint (gpt-5.2)
 
 ### Example 7: AI Summary Generation
 
@@ -462,7 +462,7 @@ export default App;
 ```
 
 **Key Features:**
-- **Backend:** OpenAI (gpt-4o-mini, gpt-4o, gpt-4.1)
+- **Backend:** OpenAI (gpt-5.2, gpt-4o, gpt-4.1)
 - **Performance:** ~3-8s for 50 checkpoints
 - **Chunking:** Automatic chunking based on token count (single/chunk/tiny strategies)
 - **Caching:** Persistent IndexedDB cache (reuses summaries across sessions)
@@ -525,7 +525,7 @@ export default App;
 **Performance Tips:**
 - Use `mode: 'fast'` for translation if you prioritize speed over completeness
 - Enable caching to avoid re-translation on subsequent analyses
-- Use gpt-4o-mini for cost efficiency, gpt-4o for best quality
+- Use gpt-5.2 for optimal balance of speed, cost, and quality
 
 ### Example 9: AI Settings UI (Advanced)
 
@@ -608,7 +608,7 @@ export default App;
 **UI Features:**
 - **Settings Dialog:** Click gear icon in sidebar header to open AI settings
 - **Real-time Toggle:** Enable/disable translation and summary without reloading
-- **Model Selection:** Choose between OpenAI models (gpt-4o-mini, gpt-4o, gpt-4.1)
+- **Model Selection:** Choose between OpenAI models (gpt-5.2, gpt-4o, gpt-4.1)
 - **Language Selection:** Change target language on the fly
 - **Persistent Settings:** All settings saved to localStorage
 

package/data/docs/TROUBLESHOOTING.md

@@ -122,7 +122,7 @@ config={{
     translation: {
         enabled: true,              // MUST be true
         apiKey: 'sk-...',           // Valid OpenAI API key
-        model: 'gpt-4o-mini',       // Correct model name
+        model: 'gpt-5.2',           // Correct model name
         targetLanguage: 'de',       // ISO 639-1 code
         mode: 'fast',
     },
@@ -144,14 +144,14 @@ console.log('OpenAI API Key Set:', !!localStorage.getItem('dqm_openai_apiKey'));
 **A:** Try these optimizations:
 
 1. **Switch to Fast Mode**: `mode: 'fast'` (15s timeout vs 120s for `'full'`)
-2. **Use a faster model**: `gpt-4o-mini` is optimized for speed
+2. **Use a faster model**: `gpt-5.2` is optimized for speed and cost-efficiency
 
 ```typescript
 // Optimize for speed
 translation: {
     enabled: true,
     apiKey: 'sk-...',
-    model: 'gpt-4o-mini',       // Fast + cheap
+    model: 'gpt-5.2',             // Fast + cheap
     targetLanguage: 'de',
     mode: 'fast',               // 15s timeout
 }
@@ -164,7 +164,7 @@ translation: {
 **A:** Check these:
 
 1. **OpenAI API Key**: Summary requires valid OpenAI key
-2. **Model**: Ensure model supports JSON mode (`gpt-4o-mini`, `gpt-4o`, `gpt-4.1`)
+2. **Model**: Ensure model supports JSON mode (`gpt-5.2`, `gpt-4o`, `gpt-4.1`)
 3. **Timeout**: Increase timeout if analysis has many checkpoints
 
 ```typescript

package/data/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@crownpeak/dqm-react-component",
-  "version": "1.2.4",
+  "version": "1.3.0",
   "private": false,
   "description": "A React component for Crownpeak Digital Quality Management (DQM) integration",
   "type": "module",

package/data/website/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@crownpeak/dqm-react-component-website",
-  "version": "1.2.4",
+  "version": "1.3.0",
   "private": true,
   "scripts": {
     "dev": "next dev",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@crownpeak/dqm-react-component-dev-mcp",
-  "version": "1.2.4",
+  "version": "1.3.0",
   "description": "MCP Server for Crownpeak DQM React Component documentation - powered by Probe",
   "author": "Crownpeak Technology GmbH",
   "license": "MIT",