-
Notifications
You must be signed in to change notification settings - Fork 4.4k
Description
Proposal: Donate the SharpAIKit Project to Microsoft
Hello Microsoft Team,
My name is Dustin Dong, the author and current maintainer of SharpAIKit, an advanced AI application framework for .NET.
I would like to formally propose transferring SharpAIKit to Microsoft as an official .NET / Azure AI developer framework.
📘 What is SharpAIKit?
SharpAIKit is a modern AI framework for .NET that provides:
✔ FSM-based AI graph orchestration
(Comparable to LangGraph / DSPy workflows, but fully native to .NET)
✔ Native C# Code Interpreter (Roslyn)
A safe, high-performance, Python-free code interpreter for AI agents.
✔ DSPy-style Prompt Optimizer
Self-improving LLM pipelines fully integrated into .NET.
✔ Unified LLM Client
Supports OpenAI, Azure OpenAI, DeepSeek, Qwen, Mistral, Yi, Groq, and all OpenAI-compatible APIs.
✔ RAG + Agents
Built-in RAG engine, memory modules, tool calling, and multi-agent support.
GitHub: https://github.com/dxpython/SharpAIKit
License: MIT
📌 Why transfer SharpAIKit to Microsoft?
1. Strengthens the .NET AI ecosystem
Currently, .NET lacks a native equivalent to LangChain/LangGraph.
SharpAIKit fills this gap completely.
2. Strong synergy with Azure OpenAI and Semantic Kernel
SharpAIKit can serve as the orchestration/runtime layer for AI developers on Azure.
3. Industry-level architecture
The FSM engine, C# interpreter, and optimizer provide a strong foundation for production AI workflows.
4. I am willing to transfer the full copyright to Microsoft
Under Microsoft’s contribution agreement.
📌 What I am proposing
✔ Transfer the full SharpAIKit project to Microsoft
or
✔ Migrate the repository under github.com/microsoft
and
✔ Continue to serve as an official maintainer or consultant
I am ready to provide:
- Project donation proposal (PDF)
- Technical architecture brief
- Governance/maintenance plan
- Copyright transfer declaration
- Migration steps
📩 Contact
Email: anonymouslook5@outlook.com
GitHub: dxpython
Please let me know how we can proceed.
Thank you!
— Dustin Dong