SkyAI

SkyAI is a Valour.gg bot powered by a self-hosted Large Language Model via Open WebUI.

It supports per-channel conversational memory and is designed for self-hosted deployments.

Features

Per-channel AI conversation history
Self-hosted LLM support (Open WebUI + Ollama)
Built with .NET
Open-source under AGPL-3.0
Privacy-conscious architecture

How It Works

SkyAI connects to:

Valour.gg API
Open WebUI /api/chat/completions endpoint
A locally hosted LLM (e.g., llama3.1 via Ollama)

Conversation history is stored in memory per channel and sent with each request to maintain contextual awareness. History is automatically trimmed to prevent excessive token usage.

Data & Privacy

SkyAI stores only the minimum data required for operation.

Stored In Memory (Per Channel)

Channel ID
Message content (AI conversation context only)
Message role (user / assistant / system)

Not Stored

Direct Messages
User credentials
Email addresses
Analytics or tracking data

Conversation history resets automatically when the bot restarts.

Full privacy policy:
https://github.com/SkyJoshua/SkyAI/blob/main/PRIVACY.md

License

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).

See the LICENSE file for details:
https://github.com/SkyJoshua/SkyAI/blob/main/LICENSE

If you modify and deploy this project publicly (including as a hosted service), you must make your source code available under the same AGPL-3.0 license.

Installation

git clone https://github.com/YOUR_USERNAME/SkyAI.git
cd SkyAI
dotnet restore

All required NuGet packages are installed automatically via SkyAI.csproj.

Requirements

.NET 8+
Open WebUI running
Ollama running with a model (e.g., llama3.1)
Valid Valour bot token

Configuration

Create a .env file in the root directory:

TOKEN=your-valour-bot-token
OPENWEBAPI=your-openwebui-api-key
OPENWEBURL=http://localhost:8080

Do not commit this file to version control.

Running the Bot

dotnet run

Commands

Command	Description
`s.ai <message>`	Send a prompt to the AI
`s.cm`	Clear channel conversation memory

Self-Hosting Notes

Ensure Ollama is running:

ollama serve

Ensure your model is installed:

ollama list

Contributing

Contributions are welcome. By submitting a contribution, you agree that your contributions will be licensed under AGPL-3.0.

Fork the repository
Create a feature branch
Submit a pull request