Prepared release v0.9.25 (#257)

README.md

@@ -35,6 +35,7 @@ Other News:
 
 Features we have recently released:
 
+- v0.9.25: Added [xAI](https://x.ai/) as a new provider. xAI provides their Grok models for generating content.
 - v0.9.23: Added support for OpenAI `o` models (`o1`, `o1-mini`, `o3`, etc.); added also an [ERI](https://github.com/MindWorkAI/ERI) server coding assistant as a preview feature behind the RAG feature flag. Your own ERI server can be used to gain access to, e.g., your enterprise data from within AI Studio.
 - v0.9.22: Added options for preview features; added embedding provider configuration for RAG (preview) and writer mode (experimental preview).
 - v0.9.18: Added the new Anthropic Heiku model; added Groq and Google Gemini as provider options.
@@ -42,7 +43,6 @@ Features we have recently released:
 - v0.9.16: Added workspace display options & improved the layout of the app window.
 - v0.9.15: Added the bias-of-the-day assistant. Tells you about a cognitive bias every day.
 - v0.9.13: You can use `ollama` providers secured with API keys.
-- v0.9.12: Added a job posting assistant to the business category and improved grammar & spelling check and rewrite assistants.
 
 ## What is AI Studio?
 
@@ -53,7 +53,7 @@ MindWork AI Studio is a desktop application available for macOS, Windows, and Li
 
 **Key advantages:**
 - **Free of charge**: The app is free to use, both for personal and commercial purposes.
-- **Independence**: You are not tied to any single provider. Instead, you can choose the provider that best suits their needs. Right now, we support OpenAI (GPT4o etc.), Mistral, Anthropic (Claude), Google Gemini, and self-hosted models using [llama.cpp](https://github.com/ggerganov/llama.cpp), [ollama](https://github.com/ollama/ollama), [LM Studio](https://lmstudio.ai/), [Groq](https://groq.com/), or [Fireworks](https://fireworks.ai/).
+- **Independence**: You are not tied to any single provider. Instead, you can choose the provider that best suits their needs. Right now, we support OpenAI (GPT4o etc.), Mistral, Anthropic (Claude), Google Gemini, xAI (Grok), and self-hosted models using [llama.cpp](https://github.com/ggerganov/llama.cpp), [ollama](https://github.com/ollama/ollama), [LM Studio](https://lmstudio.ai/), [Groq](https://groq.com/), or [Fireworks](https://fireworks.ai/).
 - **Unrestricted usage**: Unlike services like ChatGPT, which impose limits after intensive use, MindWork AI Studio offers unlimited usage through the providers API.
 - **Cost-effective**: You only pay for what you use, which can be cheaper than monthly subscription services like ChatGPT Plus, especially if used infrequently. But beware, here be dragons: For extremely intensive usage, the API costs can be significantly higher. Unfortunately, providers currently do not offer a way to display current costs in the app. Therefore, check your account with the respective provider to see how your costs are developing. When available, use prepaid and set a cost limit.
 - **Privacy**: The data entered into the app is not used for training by the providers since we are using the provider's API.

app/MindWork AI Studio/Chat/ChatThread.cs

@@ -97,4 +97,45 @@ public sealed record ChatThread
         logger.LogInformation(logMessage);
         return systemPromptText;
     }
+
+    /// <summary>
+    /// Removes a content block from this chat thread.
+    /// </summary>
+    /// <param name="content">The content block to remove.</param>
+    /// <param name="removeForRegenerate">Indicates whether the content block is removed for
+    /// regeneration purposes. True, when the content block is removed for regeneration purposes,
+    /// which will not remove the previous user block if it is hidden from the user.</param>
+    public void Remove(IContent content, bool removeForRegenerate = false)
+    {
+        var block = this.Blocks.FirstOrDefault(x => x.Content == content);
+        if(block is null)
+            return;
+
+        //
+        // Remove the previous user block if it is hidden from the user. Otherwise,
+        // the experience might be confusing for the user.
+        //
+        // Explanation, using the ERI assistant as an example:
+        // - The ERI assistant generates for every file a hidden user prompt.
+        // - In the UI, the user can only see the AI's responses, not the hidden user prompts.
+        // - Now, the user removes one AI response
+        // - The hidden user prompt is still there, but the user can't see it.
+        // - Since the user prompt is hidden, neither is it possible to remove nor edit it.
+        // - This method solves this issue by removing the hidden user prompt when the AI response is removed.
+        //
+        if (block.Role is ChatRole.AI && !removeForRegenerate)
+        {
+            var sortedBlocks = this.Blocks.OrderBy(x => x.Time).ToList();
+            var index = sortedBlocks.IndexOf(block);
+            if (index > 0)
+            {
+                var previousBlock = sortedBlocks[index - 1];
+                if (previousBlock.Role is ChatRole.USER && previousBlock.HideFromUser)
+                    this.Blocks.Remove(previousBlock);
+            }
+        }
+
+        // Remove the block from the chat thread:
+        this.Blocks.Remove(block);
+    }
 }

app/MindWork AI Studio/Chat/ContentBlockComponent.razor

@@ -12,7 +12,33 @@
             <MudText Typo="Typo.body1">@this.Role.ToName() (@this.Time)</MudText>
         </CardHeaderContent>
         <CardHeaderActions>
-            <MudIconButton Icon="@Icons.Material.Filled.ContentCopy" Color="Color.Default" OnClick="@this.CopyToClipboard" />
+            @if (this.IsSecondToLastBlock && this.Role is ChatRole.USER && this.EditLastUserBlockFunc is not null)
+            {
+                <MudTooltip Text="Edit" Placement="Placement.Bottom">
+                    <MudIconButton Icon="@Icons.Material.Filled.Edit" Color="Color.Default" OnClick="@this.EditLastUserBlock"/>
+                </MudTooltip>
+            }
+            @if (this.IsLastContentBlock && this.Role is ChatRole.USER && this.EditLastBlockFunc is not null)
+            {
+                <MudTooltip Text="Edit" Placement="Placement.Bottom">
+                    <MudIconButton Icon="@Icons.Material.Filled.Edit" Color="Color.Default" OnClick="@this.EditLastBlock"/>
+                </MudTooltip>
+            }
+            @if (this.IsLastContentBlock && this.Role is ChatRole.AI && this.RegenerateFunc is not null)
+            {
+                <MudTooltip Text="Regenerate" Placement="Placement.Bottom">
+                    <MudIconButton Icon="@Icons.Material.Filled.Recycling" Color="Color.Default" OnClick="@this.RegenerateBlock"/>
+                </MudTooltip>
+            }
+            @if (this.RemoveBlockFunc is not null)
+            {
+                <MudTooltip Text="Removes this block" Placement="Placement.Bottom">
+                    <MudIconButton Icon="@Icons.Material.Filled.Delete" Color="Color.Error" OnClick="@this.RemoveBlock"/>
+                </MudTooltip>
+            }
+            <MudTooltip Text="Copies the content to the clipboard" Placement="Placement.Bottom">
+                <MudIconButton Icon="@Icons.Material.Filled.ContentCopy" Color="Color.Default" OnClick="@this.CopyToClipboard"/>
+            </MudTooltip>
         </CardHeaderActions>
     </MudCardHeader>
     <MudCardContent>

app/MindWork AI Studio/Chat/ContentBlockComponent.razor.cs

@@ -41,6 +41,27 @@ public partial class ContentBlockComponent : ComponentBase
     [Parameter]
     public string Class { get; set; } = string.Empty;
     
+    [Parameter]
+    public bool IsLastContentBlock { get; set; } = false;
+    
+    [Parameter]
+    public bool IsSecondToLastBlock { get; set; } = false;
+
+    [Parameter]
+    public Func<IContent, Task>? RemoveBlockFunc { get; set; }
+    
+    [Parameter]
+    public Func<IContent, Task>? RegenerateFunc { get; set; }
+    
+    [Parameter]
+    public Func<IContent, Task>? EditLastBlockFunc { get; set; }
+    
+    [Parameter]
+    public Func<IContent, Task>? EditLastUserBlockFunc { get; set; }
+    
+    [Parameter]
+    public Func<bool> RegenerateEnabled { get; set; } = () => false;
+    
     [Inject]
     private RustService RustService { get; init; } = null!;
     
@@ -49,6 +70,9 @@ public partial class ContentBlockComponent : ComponentBase
     
     [Inject]
     private SettingsManager SettingsManager { get; init; } = null!;
+    
+    [Inject]
+    private IDialogService DialogService { get; init; } = null!;
 
     private bool HideContent { get; set; }
 
@@ -122,4 +146,66 @@ public partial class ContentBlockComponent : ComponentBase
     private string CardClasses => $"my-2 rounded-lg {this.Class}";
 
     private CodeBlockTheme CodeColorPalette => this.SettingsManager.IsDarkMode ? CodeBlockTheme.Dark : CodeBlockTheme.Default;
+    
+    private async Task RemoveBlock()
+    {
+        if (this.RemoveBlockFunc is null)
+            return;
+        
+        var remove = await this.DialogService.ShowMessageBox(
+            "Remove Message",
+            "Do you really want to remove this message?",
+            "Yes, remove it",
+            "No, keep it");
+        
+        if (remove.HasValue && remove.Value)
+            await this.RemoveBlockFunc(this.Content);
+    }
+    
+    private async Task RegenerateBlock()
+    {
+        if (this.RegenerateFunc is null)
+            return;
+        
+        if(this.Role is not ChatRole.AI)
+            return;
+        
+        var regenerate = await this.DialogService.ShowMessageBox(
+            "Regenerate Message",
+            "Do you really want to regenerate this message?",
+            "Yes, regenerate it",
+            "No, keep it");
+        
+        if (regenerate.HasValue && regenerate.Value)
+            await this.RegenerateFunc(this.Content);
+    }
+    
+    private async Task EditLastBlock()
+    {
+        if (this.EditLastBlockFunc is null)
+            return;
+        
+        if(this.Role is not ChatRole.USER)
+            return;
+        
+        await this.EditLastBlockFunc(this.Content);
+    }
+    
+    private async Task EditLastUserBlock()
+    {
+        if (this.EditLastUserBlockFunc is null)
+            return;
+        
+        if(this.Role is not ChatRole.USER)
+            return;
+        
+        var edit = await this.DialogService.ShowMessageBox(
+            "Edit Message",
+            "Do you really want to edit this message? In order to edit this message, the AI response will be deleted.",
+            "Yes, remove the AI response and edit it",
+            "No, keep it");
+        
+        if (edit.HasValue && edit.Value)
+            await this.EditLastUserBlockFunc(this.Content);
+    }
 }

app/MindWork AI Studio/Components/Changelog.Logs.cs

@@ -13,6 +13,8 @@ public partial class Changelog
     
     public static readonly Log[] LOGS = 
     [
+        new (200, "v0.9.25, build 200 (2025-01-04 18:33 UTC)", "v0.9.25.md"),
+        new (199, "v0.9.24, build 199 (2025-01-04 11:40 UTC)", "v0.9.24.md"),
         new (198, "v0.9.23, build 198 (2025-01-02 19:39 UTC)", "v0.9.23.md"),
         new (197, "v0.9.22, build 197 (2024-12-04 10:58 UTC)", "v0.9.22.md"),
         new (196, "v0.9.21, build 196 (2024-11-23 12:22 UTC)", "v0.9.21.md"),

app/MindWork AI Studio/Components/ChatComponent.razor

@@ -7,11 +7,26 @@
     <ChildContent>
         @if (this.ChatThread is not null)
         {
-            @foreach (var block in this.ChatThread.Blocks.OrderBy(n => n.Time))
+            var blocks = this.ChatThread.Blocks.OrderBy(n => n.Time).ToList();
+            for (var i = 0; i < blocks.Count; i++)
             {
+                var block = blocks[i];
+                var isLastBlock = i == blocks.Count - 1;
+                var isSecondLastBlock = i == blocks.Count - 2;
                 @if (!block.HideFromUser)
                 {
-                    <ContentBlockComponent Role="@block.Role" Type="@block.ContentType" Time="@block.Time" Content="@block.Content"/>
+                    <ContentBlockComponent
+                        Role="@block.Role"
+                        Type="@block.ContentType"
+                        Time="@block.Time"
+                        Content="@block.Content"
+                        RemoveBlockFunc="@this.RemoveBlock"
+                        IsLastContentBlock="@isLastBlock"
+                        IsSecondToLastBlock="@isSecondLastBlock"
+                        RegenerateFunc="@this.RegenerateBlock"
+                        RegenerateEnabled="@(() => this.IsProviderSelected)"
+                        EditLastBlockFunc="@this.EditLastBlock"
+                        EditLastUserBlockFunc="@this.EditLastUserBlock"/>
                 }
             }
         }
@@ -85,6 +100,13 @@
                 <ConfidenceInfo Mode="ConfidenceInfoMode.ICON" LLMProvider="@this.Provider.UsedLLMProvider"/>
             }
 
+            @if (this.isStreaming && this.cancellationTokenSource is not null)
+            {
+                <MudTooltip Text="Stop generation" Placement="@TOOLBAR_TOOLTIP_PLACEMENT">
+                    <MudIconButton Icon="@Icons.Material.Filled.Stop" Color="Color.Error" OnClick="() => this.CancelStreaming()"/>
+                </MudTooltip>
+            }
+
             <ProfileSelection CurrentProfile="@this.currentProfile" CurrentProfileChanged="@this.ProfileWasChanged"/>
         </MudToolBar>
     </FooterContent>

app/MindWork AI Studio/Components/ChatComponent.razor.cs

@@ -56,6 +56,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
     private bool autoSaveEnabled;
     private string currentWorkspaceName = string.Empty;
     private Guid currentWorkspaceId = Guid.Empty;
+    private CancellationTokenSource? cancellationTokenSource;
     
     // Unfortunately, we need the input field reference to blur the focus away. Without
     // this, we cannot clear the input field.
@@ -75,6 +76,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         if (deferredContent is not null)
         {
             this.ChatThread = deferredContent;
+            this.Logger.LogInformation($"The chat '{this.ChatThread.Name}' with {this.ChatThread.Blocks.Count} messages was deferred and will be rendered now.");
             await this.ChatThreadChanged.InvokeAsync(this.ChatThread);
             
             if (this.ChatThread is not null)
@@ -105,13 +107,13 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
                         await WorkspaceBehaviour.EnsureBiasWorkspace();
                 }
             }
-            
-            if (this.SettingsManager.ConfigurationData.Chat.ShowLatestMessageAfterLoading)
-            {
-                this.mustScrollToBottomAfterRender = true;
-                this.scrollRenderCountdown = 2;
-                this.StateHasChanged();
-            }
+        }
+        
+        if (this.SettingsManager.ConfigurationData.Chat.ShowLatestMessageAfterLoading)
+        {
+            this.mustScrollToBottomAfterRender = true;
+            this.scrollRenderCountdown = 4;
+            this.StateHasChanged();
         }
         
         var deferredLoading = MessageBus.INSTANCE.CheckDeferredMessages<LoadChat>(Event.LOAD_CHAT).FirstOrDefault();
@@ -119,6 +121,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         {
             this.loadChat = deferredLoading;
             this.mustLoadChat = true;
+            this.Logger.LogInformation($"The loading of the chat '{this.loadChat.ChatId}' was deferred and will be loaded now.");
         }
         
         await this.SelectProviderWhenLoadingChat();
@@ -143,18 +146,23 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         
         if (firstRender && this.mustLoadChat)
         {
+            this.Logger.LogInformation($"Try to load the chat '{this.loadChat.ChatId}' now.");
             this.mustLoadChat = false;
             this.ChatThread = await WorkspaceBehaviour.LoadChat(this.loadChat);
             
             if(this.ChatThread is not null)
             {
+                await this.ChatThreadChanged.InvokeAsync(this.ChatThread);
+                this.Logger.LogInformation($"The chat '{this.ChatThread!.ChatId}' with title '{this.ChatThread.Name}' ({this.ChatThread.Blocks.Count} messages) was loaded successfully.");
+                
                 this.currentWorkspaceName = await WorkspaceBehaviour.LoadWorkspaceName(this.ChatThread.WorkspaceId);
                 this.WorkspaceName(this.currentWorkspaceName);
                 await this.SelectProviderWhenLoadingChat();
             }
+            else
+                this.Logger.LogWarning($"The chat '{this.loadChat.ChatId}' could not be loaded.");
 
             this.StateHasChanged();
-            await this.ChatThreadChanged.InvokeAsync(this.ChatThread);
         }
         
         if(this.mustScrollToBottomAfterRender)
@@ -242,7 +250,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         }
     }
     
-    private async Task SendMessage()
+    private async Task SendMessage(bool reuseLastUserPrompt = false)
     {
         if (!this.IsProviderSelected)
             return;
@@ -252,8 +260,6 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         await this.inputField.BlurAsync();
         
         // Create a new chat thread if necessary:
-        var threadName = this.ExtractThreadName(this.userInput);
-
         if (this.ChatThread is null)
         {
             this.ChatThread = new()
@@ -263,7 +269,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
                 SystemPrompt = SystemPrompts.DEFAULT,
                 WorkspaceId = this.currentWorkspaceId,
                 ChatId = Guid.NewGuid(),
-                Name = threadName,
+                Name = this.ExtractThreadName(this.userInput),
                 Seed = this.RNG.Next(),
                 Blocks = [],
             };
@@ -274,34 +280,37 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         {
             // Set the thread name if it is empty:
             if (string.IsNullOrWhiteSpace(this.ChatThread.Name))
-                this.ChatThread.Name = threadName;
+                this.ChatThread.Name = this.ExtractThreadName(this.userInput);
             
             // Update provider and profile:
             this.ChatThread.SelectedProvider = this.Provider.Id;
             this.ChatThread.SelectedProfile = this.currentProfile.Id;
         }
-        
-        //
-        // Add the user message to the thread:
-        //
-        var time = DateTimeOffset.Now;
-        this.ChatThread?.Blocks.Add(new ContentBlock
-        {
-            Time = time,
-            ContentType = ContentType.TEXT,
-            Role = ChatRole.USER,
-            Content = new ContentText
-            {
-                Text = this.userInput,
-            },
-        });
 
-        // Save the chat:
-        if (this.SettingsManager.ConfigurationData.Workspace.StorageBehavior is WorkspaceStorageBehavior.STORE_CHATS_AUTOMATICALLY)
+        var time = DateTimeOffset.Now;
+        if (!reuseLastUserPrompt)
         {
-            await this.SaveThread();
-            this.hasUnsavedChanges = false;
-            this.StateHasChanged();
+            //
+            // Add the user message to the thread:
+            //
+            this.ChatThread?.Blocks.Add(new ContentBlock
+            {
+                Time = time,
+                ContentType = ContentType.TEXT,
+                Role = ChatRole.USER,
+                Content = new ContentText
+                {
+                    Text = this.userInput,
+                },
+            });
+
+            // Save the chat:
+            if (this.SettingsManager.ConfigurationData.Workspace.StorageBehavior is WorkspaceStorageBehavior.STORE_CHATS_AUTOMATICALLY)
+            {
+                await this.SaveThread();
+                this.hasUnsavedChanges = false;
+                this.StateHasChanged();
+            }
         }
 
         //
@@ -335,15 +344,20 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
             this.scrollRenderCountdown = 2;
         }
         
-        this.StateHasChanged();
-        
         this.Logger.LogDebug($"Start processing user input using provider '{this.Provider.InstanceName}' with model '{this.Provider.Model}'.");
         
-        // Use the selected provider to get the AI response.
-        // By awaiting this line, we wait for the entire
-        // content to be streamed.
-        await aiText.CreateFromProviderAsync(this.Provider.CreateProvider(this.Logger), this.SettingsManager, this.Provider.Model, this.ChatThread);
+        using (this.cancellationTokenSource = new())
+        {
+            this.StateHasChanged();
+            
+            // Use the selected provider to get the AI response.
+            // By awaiting this line, we wait for the entire
+            // content to be streamed.
+            await aiText.CreateFromProviderAsync(this.Provider.CreateProvider(this.Logger), this.SettingsManager, this.Provider.Model, this.ChatThread, this.cancellationTokenSource.Token);
+        }
         
+        this.cancellationTokenSource = null;
+
         // Save the chat:
         if (this.SettingsManager.ConfigurationData.Workspace.StorageBehavior is WorkspaceStorageBehavior.STORE_CHATS_AUTOMATICALLY)
         {
@@ -356,6 +370,13 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         this.StateHasChanged();
     }
     
+    private async Task CancelStreaming()
+    {
+        if (this.cancellationTokenSource is not null)
+            if(!this.cancellationTokenSource.IsCancellationRequested)
+                await this.cancellationTokenSource.CancelAsync();
+    }
+    
     private async Task SaveThread()
     {
         if(this.ChatThread is null)
@@ -509,8 +530,6 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         this.WorkspaceName(this.currentWorkspaceName);
         
         await this.SelectProviderWhenLoadingChat();
-        
-        this.userInput = string.Empty;
         if (this.SettingsManager.ConfigurationData.Chat.ShowLatestMessageAfterLoading)
         {
             this.mustScrollToBottomAfterRender = true;
@@ -537,7 +556,7 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
     {
         var chatProvider = this.ChatThread?.SelectedProvider;
         var chatProfile = this.ChatThread?.SelectedProfile;
-        
+
         switch (this.SettingsManager.ConfigurationData.Chat.LoadingProviderBehavior)
         {
             default:
@@ -571,6 +590,67 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
         await MessageBus.INSTANCE.SendMessage<bool>(this, Event.WORKSPACE_TOGGLE_OVERLAY);
     }
     
+    private async Task RemoveBlock(IContent block)
+    {
+        if(this.ChatThread is null)
+            return;
+        
+        this.ChatThread.Remove(block);
+        this.hasUnsavedChanges = true;
+        await this.SaveThread();
+        this.StateHasChanged();
+    }
+
+    private async Task RegenerateBlock(IContent aiBlock)
+    {
+        if(this.ChatThread is null)
+            return;
+        
+        this.ChatThread.Remove(aiBlock, removeForRegenerate: true);
+        this.hasUnsavedChanges = true;
+        this.StateHasChanged();
+        
+        await this.SendMessage(reuseLastUserPrompt: true);
+    }
+    
+    private Task EditLastUserBlock(IContent block)
+    {
+        if(this.ChatThread is null)
+            return Task.CompletedTask;
+        
+        if (block is not ContentText textBlock)
+            return Task.CompletedTask;
+        
+        var lastBlock = this.ChatThread.Blocks.Last();
+        var lastBlockContent = lastBlock.Content;
+        if(lastBlockContent is null)
+            return Task.CompletedTask;
+        
+        this.userInput = textBlock.Text;
+        this.ChatThread.Remove(block);
+        this.ChatThread.Remove(lastBlockContent);
+        this.hasUnsavedChanges = true;
+        this.StateHasChanged();
+        
+        return Task.CompletedTask;
+    }
+    
+    private Task EditLastBlock(IContent block)
+    {
+        if(this.ChatThread is null)
+            return Task.CompletedTask;
+        
+        if (block is not ContentText textBlock)
+            return Task.CompletedTask;
+        
+        this.userInput = textBlock.Text;
+        this.ChatThread.Remove(block);
+        this.hasUnsavedChanges = true;
+        this.StateHasChanged();
+        
+        return Task.CompletedTask;
+    }
+    
     #region Overrides of MSGComponentBase
 
     public override async Task ProcessIncomingMessage<T>(ComponentBase? sendingComponent, Event triggeredEvent, T? data) where T : default
@@ -617,6 +697,14 @@ public partial class ChatComponent : MSGComponentBase, IAsyncDisposable
             await this.SaveThread();
             this.hasUnsavedChanges = false;
         }
+
+        if (this.cancellationTokenSource is not null)
+        {
+            if(!this.cancellationTokenSource.IsCancellationRequested)
+                await this.cancellationTokenSource.CancelAsync();
+            
+            this.cancellationTokenSource.Dispose();
+        }
     }
 
     #endregion

app/MindWork AI Studio/Pages/Home.razor.cs

@@ -33,7 +33,7 @@ public partial class Home : ComponentBase
     private static readonly TextItem[] ITEMS_ADVANTAGES =
     [
         new("Free of charge", "The app is free to use, both for personal and commercial purposes."),
-        new("Independence", "You are not tied to any single provider. Instead, you might choose the provider that best suits your needs. Right now, we support OpenAI (GPT4o etc.), Mistral, Anthropic (Claude), Google Gemini, and self-hosted models using llama.cpp, ollama, LM Studio, Groq, or Fireworks."),
+        new("Independence", "You are not tied to any single provider. Instead, you might choose the provider that best suits your needs. Right now, we support OpenAI (GPT4o etc.), Mistral, Anthropic (Claude), Google Gemini, xAI (Grok), and self-hosted models using llama.cpp, ollama, LM Studio, Groq, or Fireworks."),
         new("Unrestricted usage", "Unlike services like ChatGPT, which impose limits after intensive use, MindWork AI Studio offers unlimited usage through the providers API."),
         new("Cost-effective", "You only pay for what you use, which can be cheaper than monthly subscription services like ChatGPT Plus, especially if used infrequently. But beware, here be dragons: For extremely intensive usage, the API costs can be significantly higher. Unfortunately, providers currently do not offer a way to display current costs in the app. Therefore, check your account with the respective provider to see how your costs are developing. When available, use prepaid and set a cost limit."),
         new("Privacy", "The data entered into the app is not used for training by the providers since we are using the provider's API."),

app/MindWork AI Studio/Provider/Anthropic/ProviderAnthropic.cs

@@ -10,11 +10,6 @@ namespace AIStudio.Provider.Anthropic;
 
 public sealed class ProviderAnthropic(ILogger logger) : BaseProvider("https://api.anthropic.com/v1/", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-
     #region Implementation of IProvider
 
     public override string Id => LLMProviders.ANTHROPIC.ToName();
@@ -75,71 +70,9 @@ public sealed class ProviderAnthropic(ILogger logger) : BaseProvider("https://ap
             request.Content = new StringContent(chatRequest, Encoding.UTF8, "application/json");
             return request;
         }
-
-        // Send the request using exponential backoff:
-        using var responseData = await this.SendRequest(RequestBuilder, token);
-        if(responseData.IsFailedAfterAllRetries)
-        {
-            this.logger.LogError($"Anthropic chat completion failed: {responseData.ErrorMessage}");
-            yield break;
-        }
         
-        // Open the response stream:
-        var stream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-        
-        // Add a stream reader to read the stream, line by line:
-        var streamReader = new StreamReader(stream);
-        
-        // Read the stream, line by line:
-        while(!streamReader.EndOfStream)
-        {
-            // Check if the token is canceled:
-            if(token.IsCancellationRequested)
-                yield break;
-            
-            // Read the next line:
-            var line = await streamReader.ReadLineAsync(token);
-            
-            // Skip empty lines:
-            if(string.IsNullOrWhiteSpace(line))
-                continue;
-            
-            // Check for the end of the stream:
-            if(line.StartsWith("event: message_stop", StringComparison.InvariantCulture))
-                yield break;
-            
-            // Skip lines that do not start with "data: ". Regard
-            // to the specification, we only want to read the data lines:
-            if(!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                continue;
-            
-            // Ignore any type except "content_block_delta":
-            if(!line.Contains("\"content_block_delta\"", StringComparison.InvariantCulture))
-                continue;
-
-            ResponseStreamLine anthropicResponse;
-            try
-            {
-                // We know that the line starts with "data: ". Hence, we can
-                // skip the first 6 characters to get the JSON data after that.
-                var jsonData = line[6..];
-                
-                // Deserialize the JSON data:
-                anthropicResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-            }
-            catch
-            {
-                // Skip invalid JSON data:
-                continue;
-            }    
-            
-            // Skip empty responses:
-            if(anthropicResponse == default || string.IsNullOrWhiteSpace(anthropicResponse.Delta.Text))
-                continue;
-            
-            // Yield the response:
-            yield return anthropicResponse.Delta.Text;
-        }
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("Anthropic", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/Anthropic/ResponseStreamLine.cs

@@ -7,7 +7,14 @@ namespace AIStudio.Provider.Anthropic;
 /// <param name="Type">The type of the response line.</param>
 /// <param name="Index">The index of the response line.</param>
 /// <param name="Delta">The delta of the response line.</param>
-public readonly record struct ResponseStreamLine(string Type, int Index, Delta Delta);
+public readonly record struct ResponseStreamLine(string Type, int Index, Delta Delta) : IResponseStreamLine
+{
+    /// <inheritdoc />
+    public bool ContainsContent() => this != default && !string.IsNullOrWhiteSpace(this.Delta.Text);
+
+    /// <inheritdoc />
+    public string GetContent() => this.Delta.Text;
+}
 
 /// <summary>
 /// The delta object of a response line.

app/MindWork AI Studio/Provider/BaseProvider.cs

@@ -1,4 +1,6 @@
 using System.Net;
+using System.Runtime.CompilerServices;
+using System.Text.Json;
 
 using AIStudio.Chat;
 using AIStudio.Settings;
@@ -31,6 +33,11 @@ public abstract class BaseProvider : IProvider, ISecretId
     protected static readonly RustService RUST_SERVICE;
     
     protected static readonly Encryption ENCRYPTION;
+    
+    protected static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
+    {
+        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
+    };
 
     /// <summary>
     /// Constructor for the base provider.
@@ -127,4 +134,105 @@ public abstract class BaseProvider : IProvider, ISecretId
         
         return new HttpRateLimitedStreamResult(true, false, string.Empty, response);
     }
+
+    protected async IAsyncEnumerable<string> StreamChatCompletionInternal<T>(string providerName, Func<Task<HttpRequestMessage>> requestBuilder, [EnumeratorCancellation] CancellationToken token = default) where T : struct, IResponseStreamLine
+    {
+        StreamReader? streamReader = null;
+        try
+        {
+            // Send the request using exponential backoff:
+            var responseData = await this.SendRequest(requestBuilder, token);
+            if(responseData.IsFailedAfterAllRetries)
+            {
+                this.logger.LogError($"The {providerName} chat completion failed: {responseData.ErrorMessage}");
+                yield break;
+            }
+            
+            // Open the response stream:
+            var providerStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
+
+            // Add a stream reader to read the stream, line by line:
+            streamReader = new StreamReader(providerStream);
+        }
+        catch(Exception e)
+        {
+            this.logger.LogError($"Failed to stream chat completion from {providerName} '{this.InstanceName}': {e.Message}");
+        }
+
+        if (streamReader is null)
+            yield break;
+        
+        // Read the stream, line by line:
+        while (true)
+        {
+            try
+            {
+                if(streamReader.EndOfStream)
+                    break;
+            }
+            catch (Exception e)
+            {
+                this.logger.LogWarning($"Failed to read the end-of-stream state from {providerName} '{this.InstanceName}': {e.Message}");
+                break;
+            }
+                
+            // Check if the token is canceled:
+            if (token.IsCancellationRequested)
+            {
+                this.logger.LogWarning($"The user canceled the chat completion for {providerName} '{this.InstanceName}'.");
+                streamReader.Close();
+                yield break;
+            }
+
+            // Read the next line:
+            string? line;
+            try
+            {
+                line = await streamReader.ReadLineAsync(token);
+            }
+            catch (Exception e)
+            {
+                this.logger.LogError($"Failed to read the stream from {providerName} '{this.InstanceName}': {e.Message}");
+                break;
+            }
+
+            // Skip empty lines:
+            if (string.IsNullOrWhiteSpace(line))
+                continue;
+            
+            // Skip lines that do not start with "data: ". Regard
+            // to the specification, we only want to read the data lines:
+            if (!line.StartsWith("data: ", StringComparison.InvariantCulture))
+                continue;
+
+            // Check if the line is the end of the stream:
+            if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
+                yield break;
+
+            T providerResponse;
+            try
+            {
+                // We know that the line starts with "data: ". Hence, we can
+                // skip the first 6 characters to get the JSON data after that.
+                var jsonData = line[6..];
+
+                // Deserialize the JSON data:
+                providerResponse = JsonSerializer.Deserialize<T>(jsonData, JSON_SERIALIZER_OPTIONS);
+            }
+            catch
+            {
+                // Skip invalid JSON data:
+                continue;
+            }
+
+            // Skip empty responses:
+            if (!providerResponse.ContainsContent())
+                continue;
+
+            // Yield the response:
+            yield return providerResponse.GetContent();
+        }
+        
+        streamReader.Dispose();
+    }
 }

app/MindWork AI Studio/Provider/Fireworks/ProviderFireworks.cs

@@ -10,11 +10,6 @@ namespace AIStudio.Provider.Fireworks;
 
 public class ProviderFireworks(ILogger logger) : BaseProvider("https://api.fireworks.ai/inference/v1/", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-    
     #region Implementation of IProvider
 
     /// <inheritdoc />
@@ -81,67 +76,9 @@ public class ProviderFireworks(ILogger logger) : BaseProvider("https://api.firew
             request.Content = new StringContent(fireworksChatRequest, Encoding.UTF8, "application/json");
             return request;
         }
-
-        // Send the request using exponential backoff:
-        using var responseData = await this.SendRequest(RequestBuilder, token);
-        if(responseData.IsFailedAfterAllRetries)
-        {
-            this.logger.LogError($"Fireworks chat completion failed: {responseData.ErrorMessage}");
-            yield break;
-        }
         
-        // Open the response stream:
-        var fireworksStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-        
-        // Add a stream reader to read the stream, line by line:
-        var streamReader = new StreamReader(fireworksStream);
-        
-        // Read the stream, line by line:
-        while(!streamReader.EndOfStream)
-        {
-            // Check if the token is canceled:
-            if(token.IsCancellationRequested)
-                yield break;
-            
-            // Read the next line:
-            var line = await streamReader.ReadLineAsync(token);
-            
-            // Skip empty lines:
-            if(string.IsNullOrWhiteSpace(line))
-                continue;
-            
-            // Skip lines that do not start with "data: ". Regard
-            // to the specification, we only want to read the data lines:
-            if(!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                continue;
-
-            // Check if the line is the end of the stream:
-            if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
-                yield break;
-
-            ResponseStreamLine fireworksResponse;
-            try
-            {
-                // We know that the line starts with "data: ". Hence, we can
-                // skip the first 6 characters to get the JSON data after that.
-                var jsonData = line[6..];
-                
-                // Deserialize the JSON data:
-                fireworksResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-            }
-            catch
-            {
-                // Skip invalid JSON data:
-                continue;
-            }    
-            
-            // Skip empty responses:
-            if(fireworksResponse == default || fireworksResponse.Choices.Count == 0)
-                continue;
-            
-            // Yield the response:
-            yield return fireworksResponse.Choices[0].Delta.Content;
-        }
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("Fireworks", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/Fireworks/ResponseStreamLine.cs

@@ -8,7 +8,14 @@ namespace AIStudio.Provider.Fireworks;
 /// <param name="Created">The timestamp of the response.</param>
 /// <param name="Model">The model used for the response.</param>
 /// <param name="Choices">The choices made by the AI.</param>
-public readonly record struct ResponseStreamLine(string Id, string Object, uint Created, string Model, IList<Choice> Choices);
+public readonly record struct ResponseStreamLine(string Id, string Object, uint Created, string Model, IList<Choice> Choices) : IResponseStreamLine
+{
+    /// <inheritdoc />
+    public bool ContainsContent() => this != default && this.Choices.Count > 0;
+
+    /// <inheritdoc />
+    public string GetContent() => this.Choices[0].Delta.Content;
+}
 
 /// <summary>
 /// Data model for a choice made by the AI.

app/MindWork AI Studio/Provider/Google/ProviderGoogle.cs

@@ -11,11 +11,6 @@ namespace AIStudio.Provider.Google;
 
 public class ProviderGoogle(ILogger logger) : BaseProvider("https://generativelanguage.googleapis.com/v1beta/", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-    
     #region Implementation of IProvider
 
     /// <inheritdoc />
@@ -82,67 +77,9 @@ public class ProviderGoogle(ILogger logger) : BaseProvider("https://generativela
             request.Content = new StringContent(geminiChatRequest, Encoding.UTF8, "application/json");
             return request;
         }
-
-        // Send the request using exponential backoff:
-        using var responseData = await this.SendRequest(RequestBuilder, token);
-        if(responseData.IsFailedAfterAllRetries)
-        {
-            this.logger.LogError($"Google chat completion failed: {responseData.ErrorMessage}");
-            yield break;
-        }
         
-        // Open the response stream:
-        var geminiStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-        
-        // Add a stream reader to read the stream, line by line:
-        var streamReader = new StreamReader(geminiStream);
-        
-        // Read the stream, line by line:
-        while(!streamReader.EndOfStream)
-        {
-            // Check if the token is canceled:
-            if(token.IsCancellationRequested)
-                yield break;
-            
-            // Read the next line:
-            var line = await streamReader.ReadLineAsync(token);
-            
-            // Skip empty lines:
-            if(string.IsNullOrWhiteSpace(line))
-                continue;
-            
-            // Skip lines that do not start with "data: ". Regard
-            // to the specification, we only want to read the data lines:
-            if(!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                continue;
-
-            // Check if the line is the end of the stream:
-            if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
-                yield break;
-
-            ResponseStreamLine geminiResponse;
-            try
-            {
-                // We know that the line starts with "data: ". Hence, we can
-                // skip the first 6 characters to get the JSON data after that.
-                var jsonData = line[6..];
-                
-                // Deserialize the JSON data:
-                geminiResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-            }
-            catch
-            {
-                // Skip invalid JSON data:
-                continue;
-            }    
-            
-            // Skip empty responses:
-            if(geminiResponse == default || geminiResponse.Choices.Count == 0)
-                continue;
-            
-            // Yield the response:
-            yield return geminiResponse.Choices[0].Delta.Content;
-        }
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("Google", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/Groq/ProviderGroq.cs

@@ -11,11 +11,6 @@ namespace AIStudio.Provider.Groq;
 
 public class ProviderGroq(ILogger logger) : BaseProvider("https://api.groq.com/openai/v1/", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-    
     #region Implementation of IProvider
 
     /// <inheritdoc />
@@ -84,67 +79,9 @@ public class ProviderGroq(ILogger logger) : BaseProvider("https://api.groq.com/o
             request.Content = new StringContent(groqChatRequest, Encoding.UTF8, "application/json");
             return request;
         }
-
-        // Send the request using exponential backoff:
-        using var responseData = await this.SendRequest(RequestBuilder, token);
-        if(responseData.IsFailedAfterAllRetries)
-        {
-            this.logger.LogError($"Groq chat completion failed: {responseData.ErrorMessage}");
-            yield break;
-        }
         
-        // Open the response stream:
-        var groqStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-        
-        // Add a stream reader to read the stream, line by line:
-        var streamReader = new StreamReader(groqStream);
-        
-        // Read the stream, line by line:
-        while(!streamReader.EndOfStream)
-        {
-            // Check if the token is canceled:
-            if(token.IsCancellationRequested)
-                yield break;
-            
-            // Read the next line:
-            var line = await streamReader.ReadLineAsync(token);
-            
-            // Skip empty lines:
-            if(string.IsNullOrWhiteSpace(line))
-                continue;
-            
-            // Skip lines that do not start with "data: ". Regard
-            // to the specification, we only want to read the data lines:
-            if(!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                continue;
-
-            // Check if the line is the end of the stream:
-            if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
-                yield break;
-
-            ResponseStreamLine groqResponse;
-            try
-            {
-                // We know that the line starts with "data: ". Hence, we can
-                // skip the first 6 characters to get the JSON data after that.
-                var jsonData = line[6..];
-                
-                // Deserialize the JSON data:
-                groqResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-            }
-            catch
-            {
-                // Skip invalid JSON data:
-                continue;
-            }    
-            
-            // Skip empty responses:
-            if(groqResponse == default || groqResponse.Choices.Count == 0)
-                continue;
-            
-            // Yield the response:
-            yield return groqResponse.Choices[0].Delta.Content;
-        }
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("Groq", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/IResponseStreamLine.cs

@@ -0,0 +1,16 @@
+namespace AIStudio.Provider;
+
+public interface IResponseStreamLine
+{
+    /// <summary>
+    /// Checks if the response line contains any content.
+    /// </summary>
+    /// <returns>True when the response line contains content, false otherwise.</returns>
+    public bool ContainsContent();
+    
+    /// <summary>
+    /// Gets the content of the response line.
+    /// </summary>
+    /// <returns>The content of the response line.</returns>
+    public string GetContent();
+}

app/MindWork AI Studio/Provider/LLMProviders.cs

@@ -11,6 +11,7 @@ public enum LLMProviders
     ANTHROPIC = 2,
     MISTRAL = 3,
     GOOGLE = 7,
+    X = 8,
     
     FIREWORKS = 5,
     GROQ = 6,

app/MindWork AI Studio/Provider/LLMProvidersExtensions.cs

@@ -5,6 +5,7 @@ using AIStudio.Provider.Groq;
 using AIStudio.Provider.Mistral;
 using AIStudio.Provider.OpenAI;
 using AIStudio.Provider.SelfHosted;
+using AIStudio.Provider.X;
 using AIStudio.Settings;
 
 using Host = AIStudio.Provider.SelfHosted.Host;
@@ -26,6 +27,7 @@ public static class LLMProvidersExtensions
         LLMProviders.ANTHROPIC => "Anthropic",
         LLMProviders.MISTRAL => "Mistral",
         LLMProviders.GOOGLE => "Google",
+        LLMProviders.X => "xAI",
         
         LLMProviders.GROQ => "Groq",
         LLMProviders.FIREWORKS => "Fireworks.ai",
@@ -62,6 +64,8 @@ public static class LLMProvidersExtensions
         
         LLMProviders.MISTRAL => Confidence.GDPR_NO_TRAINING.WithRegion("Europe, France").WithSources("https://mistral.ai/terms/#terms-of-service-la-plateforme").WithLevel(settingsManager.GetConfiguredConfidenceLevel(llmProvider)),
         
+        LLMProviders.X => Confidence.USA_NO_TRAINING.WithRegion("America, U.S.").WithSources("https://x.ai/legal/terms-of-service-enterprise").WithLevel(settingsManager.GetConfiguredConfidenceLevel(llmProvider)),
+        
         LLMProviders.SELF_HOSTED => Confidence.SELF_HOSTED.WithLevel(settingsManager.GetConfiguredConfidenceLevel(llmProvider)),
         
         _ => Confidence.UNKNOWN.WithLevel(settingsManager.GetConfiguredConfidenceLevel(llmProvider)),
@@ -87,6 +91,7 @@ public static class LLMProvidersExtensions
         LLMProviders.GROQ => false,
         LLMProviders.ANTHROPIC => false,
         LLMProviders.FIREWORKS => false,
+        LLMProviders.X => false,
         
         //
         // Self-hosted providers are treated as a special case anyway.
@@ -128,6 +133,7 @@ public static class LLMProvidersExtensions
                 LLMProviders.ANTHROPIC => new ProviderAnthropic(logger) { InstanceName = instanceName },
                 LLMProviders.MISTRAL => new ProviderMistral(logger) { InstanceName = instanceName },
                 LLMProviders.GOOGLE => new ProviderGoogle(logger) { InstanceName = instanceName },
+                LLMProviders.X => new ProviderX(logger) { InstanceName = instanceName },
                 
                 LLMProviders.GROQ => new ProviderGroq(logger) { InstanceName = instanceName },
                 LLMProviders.FIREWORKS => new ProviderFireworks(logger) { InstanceName = instanceName },
@@ -150,6 +156,7 @@ public static class LLMProvidersExtensions
         LLMProviders.MISTRAL => "https://console.mistral.ai/",
         LLMProviders.ANTHROPIC => "https://console.anthropic.com/dashboard",
         LLMProviders.GOOGLE => "https://console.cloud.google.com/",
+        LLMProviders.X => "https://accounts.x.ai/sign-up",
      
         LLMProviders.GROQ => "https://console.groq.com/",
         LLMProviders.FIREWORKS => "https://fireworks.ai/login",
@@ -162,6 +169,7 @@ public static class LLMProvidersExtensions
         LLMProviders.OPEN_AI => "https://platform.openai.com/usage",
         LLMProviders.MISTRAL => "https://console.mistral.ai/usage/",
         LLMProviders.ANTHROPIC => "https://console.anthropic.com/settings/plans",
+        LLMProviders.X => "https://console.x.ai/",
         LLMProviders.GROQ => "https://console.groq.com/settings/usage",
         LLMProviders.GOOGLE => "https://console.cloud.google.com/billing",
         LLMProviders.FIREWORKS => "https://fireworks.ai/account/billing",
@@ -174,6 +182,7 @@ public static class LLMProvidersExtensions
         LLMProviders.OPEN_AI => true,
         LLMProviders.MISTRAL => true,
         LLMProviders.ANTHROPIC => true,
+        LLMProviders.X => true,
         LLMProviders.GROQ => true,
         LLMProviders.FIREWORKS => true,
         LLMProviders.GOOGLE => true,
@@ -217,6 +226,7 @@ public static class LLMProvidersExtensions
         LLMProviders.MISTRAL => true,
         LLMProviders.ANTHROPIC => true,
         LLMProviders.GOOGLE => true,
+        LLMProviders.X => true,
         
         LLMProviders.GROQ => true,
         LLMProviders.FIREWORKS => true,
@@ -232,6 +242,7 @@ public static class LLMProvidersExtensions
         LLMProviders.MISTRAL => true,
         LLMProviders.ANTHROPIC => true,
         LLMProviders.GOOGLE => true,
+        LLMProviders.X => true,
         
         LLMProviders.GROQ => true,
         LLMProviders.FIREWORKS => true,

app/MindWork AI Studio/Provider/Mistral/ProviderMistral.cs

@@ -11,11 +11,6 @@ namespace AIStudio.Provider.Mistral;
 
 public sealed class ProviderMistral(ILogger logger) : BaseProvider("https://api.mistral.ai/v1/", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-    
     #region Implementation of IProvider
 
     public override string Id => LLMProviders.MISTRAL.ToName();
@@ -83,67 +78,9 @@ public sealed class ProviderMistral(ILogger logger) : BaseProvider("https://api.
             request.Content = new StringContent(mistralChatRequest, Encoding.UTF8, "application/json");
             return request;
         }
-
-        // Send the request using exponential backoff:
-        using var responseData = await this.SendRequest(RequestBuilder, token);
-        if(responseData.IsFailedAfterAllRetries)
-        {
-            this.logger.LogError($"Mistral chat completion failed: {responseData.ErrorMessage}");
-            yield break;
-        }
         
-        // Open the response stream:
-        var mistralStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-        
-        // Add a stream reader to read the stream, line by line:
-        var streamReader = new StreamReader(mistralStream);
-        
-        // Read the stream, line by line:
-        while(!streamReader.EndOfStream)
-        {
-            // Check if the token is canceled:
-            if(token.IsCancellationRequested)
-                yield break;
-            
-            // Read the next line:
-            var line = await streamReader.ReadLineAsync(token);
-            
-            // Skip empty lines:
-            if(string.IsNullOrWhiteSpace(line))
-                continue;
-            
-            // Skip lines that do not start with "data: ". Regard
-            // to the specification, we only want to read the data lines:
-            if(!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                continue;
-
-            // Check if the line is the end of the stream:
-            if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
-                yield break;
-
-            ResponseStreamLine mistralResponse;
-            try
-            {
-                // We know that the line starts with "data: ". Hence, we can
-                // skip the first 6 characters to get the JSON data after that.
-                var jsonData = line[6..];
-                
-                // Deserialize the JSON data:
-                mistralResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-            }
-            catch
-            {
-                // Skip invalid JSON data:
-                continue;
-            }    
-            
-            // Skip empty responses:
-            if(mistralResponse == default || mistralResponse.Choices.Count == 0)
-                continue;
-            
-            // Yield the response:
-            yield return mistralResponse.Choices[0].Delta.Content;
-        }
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("Mistral", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/OpenAI/ProviderOpenAI.cs

@@ -13,11 +13,6 @@ namespace AIStudio.Provider.OpenAI;
 /// </summary>
 public sealed class ProviderOpenAI(ILogger logger) : BaseProvider("https://api.openai.com/v1/", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-    
     #region Implementation of IProvider
 
     /// <inheritdoc />
@@ -112,66 +107,8 @@ public sealed class ProviderOpenAI(ILogger logger) : BaseProvider("https://api.o
             return request;
         }
         
-        // Send the request using exponential backoff:
-        using var responseData = await this.SendRequest(RequestBuilder, token);
-        if(responseData.IsFailedAfterAllRetries)
-        {
-            this.logger.LogError($"OpenAI chat completion failed: {responseData.ErrorMessage}");
-            yield break;
-        }
-
-        // Open the response stream:
-        var openAIStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-        
-        // Add a stream reader to read the stream, line by line:
-        var streamReader = new StreamReader(openAIStream);
-        
-        // Read the stream, line by line:
-        while(!streamReader.EndOfStream)
-        {
-            // Check if the token is canceled:
-            if(token.IsCancellationRequested)
-                yield break;
-            
-            // Read the next line:
-            var line = await streamReader.ReadLineAsync(token);
-            
-            // Skip empty lines:
-            if(string.IsNullOrWhiteSpace(line))
-                continue;
-            
-            // Skip lines that do not start with "data: ". Regard
-            // to the specification, we only want to read the data lines:
-            if(!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                continue;
-
-            // Check if the line is the end of the stream:
-            if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
-                yield break;
-
-            ResponseStreamLine openAIResponse;
-            try
-            {
-                // We know that the line starts with "data: ". Hence, we can
-                // skip the first 6 characters to get the JSON data after that.
-                var jsonData = line[6..];
-                
-                // Deserialize the JSON data:
-                openAIResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-            }
-            catch
-            {
-                // Skip invalid JSON data:
-                continue;
-            }    
-            
-            // Skip empty responses:
-            if(openAIResponse == default || openAIResponse.Choices.Count == 0)
-                continue;
-            
-            // Yield the response:
-            yield return openAIResponse.Choices[0].Delta.Content;
-        }
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("OpenAI", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/OpenAI/ResponseStreamLine.cs

@@ -9,7 +9,14 @@ namespace AIStudio.Provider.OpenAI;
 /// <param name="Model">The model used for the response.</param>
 /// <param name="SystemFingerprint">The system fingerprint; together with the seed, this allows you to reproduce the response.</param>
 /// <param name="Choices">The choices made by the AI.</param>
-public readonly record struct ResponseStreamLine(string Id, string Object, uint Created, string Model, string SystemFingerprint, IList<Choice> Choices);
+public readonly record struct ResponseStreamLine(string Id, string Object, uint Created, string Model, string SystemFingerprint, IList<Choice> Choices) : IResponseStreamLine
+{
+    /// <inheritdoc />
+    public bool ContainsContent() => this != default && this.Choices.Count > 0;
+
+    /// <inheritdoc />
+    public string GetContent() => this.Choices[0].Delta.Content;
+}
 
 /// <summary>
 /// Data model for a choice made by the AI.

app/MindWork AI Studio/Provider/SelfHosted/ProviderSelfHosted.cs

@@ -11,11 +11,6 @@ namespace AIStudio.Provider.SelfHosted;
 
 public sealed class ProviderSelfHosted(ILogger logger, Host host, string hostname) : BaseProvider($"{hostname}{host.BaseURL()}", logger)
 {
-    private static readonly JsonSerializerOptions JSON_SERIALIZER_OPTIONS = new()
-    {
-        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
-    };
-
     #region Implementation of IProvider
 
     public override string Id => LLMProviders.SELF_HOSTED.ToName();
@@ -67,91 +62,22 @@ public sealed class ProviderSelfHosted(ILogger logger, Host host, string hostnam
             MaxTokens = -1,
         }, JSON_SERIALIZER_OPTIONS);
 
-        StreamReader? streamReader = default;
-        try
+        async Task<HttpRequestMessage> RequestBuilder()
         {
-            async Task<HttpRequestMessage> RequestBuilder()
-            {
-                // Build the HTTP post request:
-                var request = new HttpRequestMessage(HttpMethod.Post, host.ChatURL());
+            // Build the HTTP post request:
+            var request = new HttpRequestMessage(HttpMethod.Post, host.ChatURL());
 
-                // Set the authorization header:
-                if (requestedSecret.Success)
-                    request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", await requestedSecret.Secret.Decrypt(ENCRYPTION));
+            // Set the authorization header:
+            if (requestedSecret.Success)
+                request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", await requestedSecret.Secret.Decrypt(ENCRYPTION));
 
-                // Set the content:
-                request.Content = new StringContent(providerChatRequest, Encoding.UTF8, "application/json");
-                return request;
-            }
-            
-            // Send the request using exponential backoff:
-            using var responseData = await this.SendRequest(RequestBuilder, token);
-            if(responseData.IsFailedAfterAllRetries)
-            {
-                this.logger.LogError($"Self-hosted provider's chat completion failed: {responseData.ErrorMessage}");
-                yield break;
-            }
-
-            // Open the response stream:
-            var providerStream = await responseData.Response!.Content.ReadAsStreamAsync(token);
-
-            // Add a stream reader to read the stream, line by line:
-            streamReader = new StreamReader(providerStream);
-        }
-        catch(Exception e)
-        {
-            this.logger.LogError($"Failed to stream chat completion from self-hosted provider '{this.InstanceName}': {e.Message}");
-        }
-
-        if (streamReader is not null)
-        {
-            // Read the stream, line by line:
-            while (!streamReader.EndOfStream)
-            {
-                // Check if the token is canceled:
-                if (token.IsCancellationRequested)
-                    yield break;
-
-                // Read the next line:
-                var line = await streamReader.ReadLineAsync(token);
-
-                // Skip empty lines:
-                if (string.IsNullOrWhiteSpace(line))
-                    continue;
-
-                // Skip lines that do not start with "data: ". Regard
-                // to the specification, we only want to read the data lines:
-                if (!line.StartsWith("data: ", StringComparison.InvariantCulture))
-                    continue;
-
-                // Check if the line is the end of the stream:
-                if (line.StartsWith("data: [DONE]", StringComparison.InvariantCulture))
-                    yield break;
-
-                ResponseStreamLine providerResponse;
-                try
-                {
-                    // We know that the line starts with "data: ". Hence, we can
-                    // skip the first 6 characters to get the JSON data after that.
-                    var jsonData = line[6..];
-
-                    // Deserialize the JSON data:
-                    providerResponse = JsonSerializer.Deserialize<ResponseStreamLine>(jsonData, JSON_SERIALIZER_OPTIONS);
-                }
-                catch
-                {
-                    // Skip invalid JSON data:
-                    continue;
-                }
-
-                // Skip empty responses:
-                if (providerResponse == default || providerResponse.Choices.Count == 0)
-                    continue;
-
-                // Yield the response:
-                yield return providerResponse.Choices[0].Delta.Content;
-            }
+            // Set the content:
+            request.Content = new StringContent(providerChatRequest, Encoding.UTF8, "application/json");
+            return request;
         }
+        
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("self-hosted provider", RequestBuilder, token))
+            yield return content;
     }
 
     #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously

app/MindWork AI Studio/Provider/X/ProviderX.cs

@@ -0,0 +1,152 @@
+using System.Net.Http.Headers;
+using System.Runtime.CompilerServices;
+using System.Text;
+using System.Text.Json;
+
+using AIStudio.Chat;
+using AIStudio.Provider.OpenAI;
+using AIStudio.Settings;
+
+namespace AIStudio.Provider.X;
+
+public sealed class ProviderX(ILogger logger) : BaseProvider("https://api.x.ai/v1/", logger)
+{
+    #region Implementation of IProvider
+
+    /// <inheritdoc />
+    public override string Id => LLMProviders.X.ToName();
+
+    /// <inheritdoc />
+    public override string InstanceName { get; set; } = "xAI";
+
+    /// <inheritdoc />
+    public override async IAsyncEnumerable<string> StreamChatCompletion(Model chatModel, ChatThread chatThread, SettingsManager settingsManager, [EnumeratorCancellation] CancellationToken token = default)
+    {
+        // Get the API key:
+        var requestedSecret = await RUST_SERVICE.GetAPIKey(this);
+        if(!requestedSecret.Success)
+            yield break;
+        
+        // Prepare the system prompt:
+        var systemPrompt = new Message
+        {
+            Role = "system",
+            Content = chatThread.PrepareSystemPrompt(settingsManager, chatThread, this.logger),
+        };
+        
+        // Prepare the xAI HTTP chat request:
+        var xChatRequest = JsonSerializer.Serialize(new ChatRequest
+        {
+            Model = chatModel.Id,
+            
+            // Build the messages:
+            // - First of all the system prompt
+            // - Then none-empty user and AI messages
+            Messages = [systemPrompt, ..chatThread.Blocks.Where(n => n.ContentType is ContentType.TEXT && !string.IsNullOrWhiteSpace((n.Content as ContentText)?.Text)).Select(n => new Message
+            {
+                Role = n.Role switch
+                {
+                    ChatRole.USER => "user",
+                    ChatRole.AI => "assistant",
+                    ChatRole.AGENT => "assistant",
+                    ChatRole.SYSTEM => "system",
+
+                    _ => "user",
+                },
+
+                Content = n.Content switch
+                {
+                    ContentText text => text.Text,
+                    _ => string.Empty,
+                }
+            }).ToList()],
+
+            Seed = chatThread.Seed,
+            
+            // Right now, we only support streaming completions:
+            Stream = true,
+        }, JSON_SERIALIZER_OPTIONS);
+
+        async Task<HttpRequestMessage> RequestBuilder()
+        {
+            // Build the HTTP post request:
+            var request = new HttpRequestMessage(HttpMethod.Post, "chat/completions");
+
+            // Set the authorization header:
+            request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", await requestedSecret.Secret.Decrypt(ENCRYPTION));
+
+            // Set the content:
+            request.Content = new StringContent(xChatRequest, Encoding.UTF8, "application/json");
+            return request;
+        }
+        
+        await foreach (var content in this.StreamChatCompletionInternal<ResponseStreamLine>("xAI", RequestBuilder, token))
+            yield return content;
+    }
+
+    #pragma warning disable CS1998 // Async method lacks 'await' operators and will run synchronously
+    /// <inheritdoc />
+    public override async IAsyncEnumerable<ImageURL> StreamImageCompletion(Model imageModel, string promptPositive, string promptNegative = FilterOperator.String.Empty, ImageURL referenceImageURL = default, [EnumeratorCancellation] CancellationToken token = default)
+    {
+        yield break;
+    }
+    #pragma warning restore CS1998 // Async method lacks 'await' operators and will run synchronously
+
+    /// <inheritdoc />
+    public override Task<IEnumerable<Model>> GetTextModels(string? apiKeyProvisional = null, CancellationToken token = default)
+    {
+        return this.LoadModels(["grok-"], token, apiKeyProvisional);
+    }
+
+    /// <inheritdoc />
+    public override Task<IEnumerable<Model>> GetImageModels(string? apiKeyProvisional = null, CancellationToken token = default)
+    {
+        return Task.FromResult<IEnumerable<Model>>([]);
+    }
+    
+    /// <inheritdoc />
+    public override Task<IEnumerable<Model>> GetEmbeddingModels(string? apiKeyProvisional = null, CancellationToken token = default)
+    {
+        return Task.FromResult<IEnumerable<Model>>([]);
+    }
+
+    #endregion
+    
+    private async Task<IEnumerable<Model>> LoadModels(string[] prefixes, CancellationToken token, string? apiKeyProvisional = null)
+    {
+        var secretKey = apiKeyProvisional switch
+        {
+            not null => apiKeyProvisional,
+            _ => await RUST_SERVICE.GetAPIKey(this) switch
+            {
+                { Success: true } result => await result.Secret.Decrypt(ENCRYPTION),
+                _ => null,
+            }
+        };
+
+        if (secretKey is null)
+            return [];
+        
+        var request = new HttpRequestMessage(HttpMethod.Get, "models");
+        request.Headers.Authorization = new AuthenticationHeaderValue("Bearer", secretKey);
+
+        var response = await this.httpClient.SendAsync(request, token);
+        if(!response.IsSuccessStatusCode)
+            return [];
+
+        var modelResponse = await response.Content.ReadFromJsonAsync<ModelsResponse>(token);
+        
+        //
+        // The API does not return the alias model names, so we have to add them manually:
+        // Right now, the only alias to add is `grok-2-latest`.
+        //
+        return modelResponse.Data.Where(model => prefixes.Any(prefix => model.Id.StartsWith(prefix, StringComparison.InvariantCulture)))
+            .Concat([
+                new Model
+                {
+                    Id = "grok-2-latest",
+                    DisplayName = "Grok 2.0 (latest)",
+                }
+            ]);
+    }
+}

app/MindWork AI Studio/Tools/HttpRateLimitedStreamResult.cs

@@ -10,14 +10,4 @@ public readonly record struct HttpRateLimitedStreamResult(
     bool IsSuccessful,
     bool IsFailedAfterAllRetries,
     string ErrorMessage,
-    HttpResponseMessage? Response) : IDisposable
-{
-    #region IDisposable
-
-    public void Dispose()
-    {
-        this.Response?.Dispose();
-    }
-
-    #endregion
-}
+    HttpResponseMessage? Response);

app/MindWork AI Studio/wwwroot/changelog/v0.9.24.md

@@ -0,0 +1,7 @@
+# v0.9.24, build 199 (2025-01-04 11:40 UTC)
+- Added a button to remove a message from the chat thread.
+- Added a button to regenerate the last AI response.
+- Added a button to edit the last user message.
+- Added a button to stop the AI from generating more tokens.
+- Improved the stream handling for all providers. The stream handling is now very resilient and handles all kinds of network issues.
+- Fixed a streaming bug that was particularly visible with self-hosted providers.

app/MindWork AI Studio/wwwroot/changelog/v0.9.25.md

@@ -0,0 +1,6 @@
+# v0.9.25, build 200 (2025-01-04 18:33 UTC)
+- Added xAI as a new provider. xAI provides their Grok models for generating content.
+- Improved the stop generation button behavior to ensure that the AI stops generating content immediately (which will save compute time, energy and financial resources).
+- Fixed the chat scrolling behavior after toggling the workspaces.
+- Fixed an issue when loading the today's bias chat, triggered by the bias assistant.
+- Restructured the streaming network code to be centralized out of the individual providers. This will allow for easier maintenance and updates in the future.

metadata.txt

@@ -1,9 +1,9 @@
-0.9.23
-2025-01-02 19:39:00 UTC
-198
+0.9.25
+2025-01-04 18:33:07 UTC
+200
 8.0.111 (commit f3835a7fb4)
 8.0.11 (commit 9cb3b725e3)
 1.83.0 (commit 90b35a623)
 7.15.0
 1.8.1
-dd03889d10b, release
+e4b1ea32fb6, release

runtime/Cargo.lock

@@ -2319,7 +2319,7 @@ checksum = "6877bb514081ee2a7ff5ef9de3281f14a4dd4bceac4c09388074a6b5df8a139a"
 
 [[package]]
 name = "mindwork-ai-studio"
-version = "0.9.23"
+version = "0.9.25"
 dependencies = [
  "aes",
  "arboard",

runtime/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "mindwork-ai-studio"
-version = "0.9.23"
+version = "0.9.25"
 edition = "2021"
 description = "MindWork AI Studio"
 authors = ["Thorsten Sommer"]

runtime/tauri.conf.json

@@ -6,7 +6,7 @@
   },
   "package": {
     "productName": "MindWork AI Studio",
-	"version": "0.9.23"
+	"version": "0.9.25"
   },
   "tauri": {
     "allowlist": {