release/0.5.0-alpha

.github/labels.yml

@@ -68,6 +68,9 @@
 - name: "component: AI TextListGenerate"
   color: "000"
   description: "Issues related to the AI Text List Generate component"
+- name: "component: AI ImgGenerate"
+  color: "000"
+  description: "Issues related to the AI Image Generate component"
 - name: "component: AI ScriptReview"
   color: "000"
   description: "Issues related to the AI Script Review component"
@@ -113,6 +116,9 @@
 - name: "component: LoadGhJSON"
   color: "000"
   description: "Issues related to the Load GhJSON file component"
+- name: "component: ImageViewer"
+  color: "000"
+  description: "Issues related to the Image Viewer component"
 
 # Provider Labels
 - name: "provider: MistralAI"

.windsurf/workflows/new-branch.md

@@ -19,7 +19,7 @@ Use the following pattern for branch names:
 Where:
 - `prefix`: Type of change (see available prefixes below)
 - `X.Y.Z`: Target version for the changes
-- `descriptive-title`: 1-4 words describing the branch purpose (kebab-case)
+- `descriptive-title`: 1-3 words describing the branch purpose (kebab-case)
 
 ### Available Prefixes
 

CHANGELOG.md

@@ -7,6 +7,47 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.5.0-alpha] - 2025-07-29
+
+### Added
+
+- **Model Capability Management System**
+  - Introduced `AIModelCapabilities` and `AIModelCapabilityRegistry` for centralized, persistent model capability tracking.
+  - Added capability checking and filtering methods for models (e.g., `GetCapabilities`, `SetCapabilities`, `FindModelsWithCapabilities`).
+  - Tool-specific capability validation now prevents execution with incompatible models.
+  - Default model is now managed by the `AIModelCapabilityRegistry`. Multiple models can be defined as Default for a set of capabilities.
+  - `AIStatefulAsyncComponentBase` will now try to use the default model if the specified model is not compatible with the tool.
+- **Provider-Specific Capability Management**
+  - MistralAI:
+    - Added `MistralModelManager` for dynamic API-based capability detection and registration.
+    - Models now update their capabilities by querying the `/v1/models/{model_id}` endpoint.
+    - Automatic mapping of Mistral model features (chat, function calling, vision) to internal capability flags.
+  - OpenAI & DeepSeek:
+    - Static mapping for capabilities, with support for function calling, structured output, and image generation.
+- **Image Generation Support**: Comprehensive AI image generation capabilities using OpenAI DALL-E models.
+  - New `DefaultImgModel` property in `IAIProvider` interface for provider capability detection.
+  - New `img_generate` AI tool with support for prompt, size, quality, and style parameters.
+  - Enhanced `AIUtils.GenerateImage()` method with provider-agnostic image generation.
+  - New `AIImgGenerateComponent` UI component in SmartHopper > Img category.
+- Improvements in `AITools`:
+  - New `includeSubcategories` parameter to `gh_list_categories` tool.
+  - New `nameFilter`, `includeDetails` and `maxResults` parameters to `gh_list_components` tool.
+  - New `ImageViewer` component to visualize output images on canvas and save them to disk.
+- Added component existence and connection type validation to `GHJsonLocal`.
+- **Settings management in AI Providers**:
+  - New `SetSetting` method in `AIProvider` that let's providers set custom settings within the provider key.
+  - New `RefreshCachedSettings` method in `AIProvider` to refresh their cached settings.
+
+### Changed
+
+- Renamed `AIProvider.InitializeSettgins` to `AIProvider.ResetCachedSettings`. Set visibility to `private`.
+
+### Fixed
+
+- `gh_put` now automatically fixes GhJSON.
+- OpenAI tool filter not being applied properly.
+- Fixed "Parsing error when output contains { }" ([#276](https://github.com/architects-toolkit/SmartHopper/issues/276)).
+
 ## [0.4.1-alpha] - 2025-07-23
 
 ### Added
@@ -53,6 +94,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Fixed certificate creation tests to handle CI environment constraints
 - Updated `GhRetrieveComponents` to use the correct ai tool `gh_list_components` instead of `gh_get_available_components`
 - Fixes "Missing required parameter: ‘response_format.json_schema' in text-list-generate with OpenAI provider" ([#259](https://github.com/architects-toolkit/SmartHopper/issues/259)).
+- Fixes "Check structured output compatibility with models" ([#273](https://github.com/architects-toolkit/SmartHopper/issues/273)).
 
 ## [0.3.6-alpha] - 2025-07-20
 

README.md

@@ -1,6 +1,6 @@
 # SmartHopper - AI-Powered Grasshopper3D Plugin
 
-[![Version](https://img.shields.io/badge/version-0%2E4%2E1--alpha-orange)](https://github.com/architects-toolkit/SmartHopper/releases)
+[![Version](https://img.shields.io/badge/version-0%2E5%2E0--alpha-orange)](https://github.com/architects-toolkit/SmartHopper/releases)
 [![Status](https://img.shields.io/badge/status-Alpha-orange)](https://github.com/architects-toolkit/SmartHopper/releases)
 [![Test results](https://img.shields.io/github/actions/workflow/status/architects-toolkit/SmartHopper/.github/workflows/ci-dotnet-tests.yml?label=.NET%20CI&logo=dotnet)](https://github.com/architects-toolkit/SmartHopper/actions/workflows/ci-dotnet-tests.yml)
 [![Grasshopper](https://img.shields.io/badge/plugin_for-Grasshopper3D-darkgreen?logo=rhinoceros)](https://www.rhino3d.com/)
@@ -71,6 +71,7 @@ After installation, all SmartHopper components will be available in the Grasshop
 | AI Text Evaluate (AiTextEvaluate)<br><sub>Return a boolean from a text content using AI-powered checks</sub> | ⚪ | 🟡 | 🟠 | 🟢 |
 | AI Text Generate (AiTextGenerate)<br><sub>Generate text content using AI</sub> | ⚪ | 🟡 | 🟠 | 🟢 |
 | AI Text List Generate (AiTextListGenerate)<br><sub>Generate lists of text content using AI</sub> | ⚪ | 🟡 | 🟠 | 🟢 |
+| AI Image Generate (AiImageGenerate)<br><sub>Generate images using AI</sub> | ⚪ | 🟡 | 🟠 | 🟢 |
 | AI Script Review (AiScriptReview)<br><sub>Make a review of a script, using AI</sub> | ⚪ | - | - | - |
 | AI Script Edit (AiScriptEdit)<br><sub>Modify an existing script using AI</sub> | ⚪ | - | - | - |
 | AI Script New (AiScriptNew)<br><sub>Generate a script using AI</sub> | ⚪ | - | - | - |
@@ -96,6 +97,7 @@ AI Tools are the interface between AI and Grasshopper, allowing to, for example,
 |-----------|-------------|:-------:|:-----------:|:-------:|:-----------:|
 | text_evaluate | Evaluates text against a true/false question | ⚪ | 🟡 | 🟠 | 🟢 |
 | text_generate | Generates text based on a prompt and optional instructions | ⚪ | 🟡 | 🟠 | 🟢 |
+| img_generate | Generates an image based on a prompt and optional instructions | ⚪ | 🟡 | 🟠 | 🟢 |
 | list_evaluate | Evaluates a list based on a natural language question | ⚪ | 🟡 | 🟠 | 🟢 |
 | list_filter | Filters a list based on natural language criteria | ⚪ | 🟡 | 🟠 | 🟢 |
 | list_generate | Generates a list based on a natural language prompt | ⚪ | 🟡 | 🟠 | 🟢 |

Solution.props

@@ -1,5 +1,5 @@
 <Project>
   <PropertyGroup>
-    <SolutionVersion>0.4.1-alpha</SolutionVersion>
+    <SolutionVersion>0.5.0-alpha</SolutionVersion>
   </PropertyGroup>
 </Project>

src/SmartHopper.Components/AI/AIChatComponent.cs

@@ -228,8 +228,8 @@ public override async System.Threading.Tasks.Task DoWorkAsync(CancellationToken
                     this.progressReporter?.Invoke("Starting web chat interface...");
 
                     // Get the actual provider name to use
-                    string actualProvider = this.component.GetActualProviderName();
-                    Debug.WriteLine($"[AIChatWorker] Using Provider: {actualProvider} (Selected: {this.component.GetActualProviderName()})");
+                    string actualProvider = this.component.GetActualAIProviderName();
+                    Debug.WriteLine($"[AIChatWorker] Using Provider: {actualProvider} (Selected: {this.component.GetActualAIProviderName()})");
 
                     // Create a web chat worker
                     var chatWorker = WebChatUtils.CreateWebChatWorker(

src/SmartHopper.Components/AI/AIModelsComponent.cs

@@ -127,7 +127,7 @@ public override async Task DoWorkAsync(CancellationToken token)
                 try
                 {
                     // Get the current AI provider
-                    var provider = this._parent.GetCurrentAIProvider();
+                    var provider = this._parent.GetActualAIProvider();
                     if (provider == null)
                     {
                         this._result["Success"] = false;
@@ -136,7 +136,7 @@ public override async Task DoWorkAsync(CancellationToken token)
                     }
 
                     // Retrieve available models
-                    var models = await provider.RetrieveAvailableModels();
+                    var models = await provider.Models.RetrieveAvailable().ConfigureAwait(false);
                     if (models == null || !models.Any())
                     {
                         this._result["Success"] = false;