Ollama Chat Interface

A modern Windows desktop chat application for interacting with Ollama LLM models. Built with .NET 8 and WPF, featuring a clean interface inspired by Claude and ChatGPT.

Features

• Modern Dark Theme UI - Clean, responsive interface similar to ChatGPT/Claude
• Multiple Chat Support - Create and manage multiple conversations
• Project Organization - Organize chats into projects and subdirectories
• File Attachments - Upload and attach files to your messages (including images for vision models)
• Artifact Storage - Automatically extracts and stores code blocks from responses
• Configurable Ollama Connection - Connect to local or remote Ollama servers
• Streaming Responses - Real-time streaming of model responses
• Model Selection - Choose from available models on your Ollama server
• Persistent Storage - All chats and artifacts stored locally in SQLite

Requirements

• Windows 10/11
• .NET 8 SDK or later
• Ollama running locally or on a remote server

Installation

From Source

1. Clone the repository:

   git clone <repository-url>
   cd chat-interface

2. Build the application:

   dotnet build OllamaChat.sln -c Release

3. Run the application:

   dotnet run --project OllamaChat/OllamaChat.csproj

Publish as Standalone

To create a self-contained executable:

dotnet publish OllamaChat/OllamaChat.csproj -c Release -r win-x64 --self-contained true -o ./publish

The executable will be in the ./publish folder.

Configuration

appsettings.json

The application can be configured via appsettings.json located in the application directory:

{
  "Ollama": {
    "BaseUrl": "http://localhost:11434",
    "DefaultModel": "llama3.2",
    "TimeoutSeconds": 300,
    "StreamResponses": true,
    "DefaultOptions": {
      "Temperature": 0.7,
      "NumCtx": 4096,
      "TopP": 0.9,
      "TopK": 40,
      "RepeatPenalty": 1.1
    }
  }
}

Configuration Options

Option	Description	Default
BaseUrl	URL of your Ollama server	http://localhost:11434
DefaultModel	Default model for new chats	llama3.2
TimeoutSeconds	Request timeout in seconds	300
StreamResponses	Enable streaming responses	true
Temperature	Model creativity (0.0-2.0)	0.7
NumCtx	Context window size	4096
TopP	Nucleus sampling	0.9
TopK	Top-K sampling	40
RepeatPenalty	Repetition penalty	1.1

Remote Ollama Server

To connect to a remote Ollama server:

1. Click the ⚙ (Settings) button in the sidebar
2. Enter the remote server URL (e.g., http://192.168.1.100:11434)
3. Click Save

Or edit appsettings.json:

{
  "Ollama": {
    "BaseUrl": "http://your-server-ip:11434"
  }
}

Note: Ensure your Ollama server is configured to accept remote connections by setting OLLAMA_HOST=0.0.0.0 on the server.

Usage

Creating a New Chat

1. Click the + New Chat button in the sidebar
2. Select a model from the dropdown in the header
3. Type your message and press Enter or click Send

Organizing Chats

• Projects: Create projects to organize related chats
• Search: Use the search box to find chats by title or content
• Recent Chats: Recently accessed chats appear in the sidebar

File Attachments

1. Click the 📎 (paperclip) button next to the input
2. Select one or more files
3. Files will be attached to your next message

Supported for Vision Models:

• Images: .png, .jpg, .jpeg, .gif, .webp

Artifacts

The application automatically extracts code blocks from assistant responses:

• View artifacts in the right panel (click 📄 to toggle)
• Copy code with the 📋 button
• Save to file with the 💾 button

Data Storage

All data is stored locally in:

• Windows: %LOCALAPPDATA%\OllamaChat\

Contents:

• ollama_chat.db - SQLite database with chats and messages
• Uploads/ - Uploaded file attachments
• Artifacts/ - Saved artifacts

Architecture

The application follows the MVVM pattern:

OllamaChat/
├── Models/           # Data models
├── ViewModels/       # View models (MVVM)
├── Views/            # WPF views (XAML)
├── Services/         # Business logic services
├── Data/             # Database context
├── Converters/       # WPF value converters
└── Resources/        # Styles and colors

Key Technologies

• .NET 8 - Application framework
• WPF - Windows UI framework
• Entity Framework Core - Database ORM
• SQLite - Local database
• CommunityToolkit.Mvvm - MVVM framework
• Markdig - Markdown rendering

Troubleshooting

"Disconnected" Status

1. Ensure Ollama is running (ollama serve)
2. Check the server URL in settings
3. Click the ↻ button to refresh connection

No Models Available

1. Install models via Ollama CLI: ollama pull llama3.2
2. Click refresh to reload available models

Slow Responses

• Reduce context length in settings
• Use a smaller model
• Ensure adequate system resources

License

MIT License - See LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit pull requests.