Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Unity MCP: AI Game Development
Learn how to integrate top LLM models with the open‑source Unity‑MCP framework, enabling AI‑driven game development using real‑time tool interaction for developers.
I am a developer of open sourced Unity-MCP project: https://github.com/IvanMurzak/Unity-MCP
I will show and explain how to use top LLM models such as ChatGPT to let them to utilize the best tools that humanity build for ourself. LLM natively understands the tools, and may use the toolset in a very similar way as a human uses tools to make something. I will cover in details the game development aspect of the topic, but briefly will talk in general about the potential of MCP (Model Context Protocol).
Unity-MCP enables LLM-driven Unity Editor automation via MCP protocol.
- UnityUnity: The cross-platform, real-time 3D development platform for building 2D, 3D, AR, and VR experiences across over 25 target platforms.Unity is the industry-standard real-time development engine, powering everything from AAA games like *Cuphead* and *Among Us* to high-fidelity simulations for film, automotive, and architecture (AEC). It employs a component-based architecture, with C# as the primary scripting language, allowing developers to 'Build Once, Deploy ANYWHERE' (mobile, desktop, console, WebGL). The platform’s versatility and extensive Asset Store make it the go-to environment for rapid prototyping and production, dominating the market for augmented and virtual reality content creation.
- C#C# is a modern, object-oriented, and type-safe programming language: it's Microsoft's flagship for development on the cross-platform .NET framework.C# (pronounced 'See-Sharp') is a powerful, open-source language developed by Microsoft for the .NET platform. It delivers a robust, object-oriented foundation with strong typing and automatic memory management (via the Garbage Collector), ensuring high stability and performance. Currently at version C# 12 (paired with .NET 8), it is the primary choice for building enterprise-grade applications: this includes web services (ASP.NET Core), cloud-native solutions (Azure), desktop apps (WPF, Windows Forms), and cross-platform mobile interfaces (.NET MAUI). Developers leverage its extensive library ecosystem and Visual Studio integration for rapid, efficient development cycles.
- ReflectionThe ability for a program to inspect, introspect, and modify its own structure (classes, methods, fields) at runtime.Reflection is a powerful metaprogramming technique: it grants a running program the capability to examine and manipulate its own structure and behavior. In languages like Java and C#, this means you can dynamically inspect classes, invoke methods, and instantiate objects, all without compile-time knowledge. This dynamic capability is critical for building flexible, generic software libraries: consider its use in serialization, dependency injection frameworks, and Object-Relational Mappers (ORMs). Use it to bypass member accessibility rules or to implement runtime configuration based on metadata. Be aware: this power comes with performance overhead and potential security risks.
- SignalRSignalR is the open-source library for ASP.NET Core that enables instant, bidirectional server-to-client content push, abstracting real-time communication protocols.SignalR establishes a persistent, full-duplex connection between server and client, moving beyond the classic HTTP request-response model. It automatically manages transport fallback, prioritizing WebSockets (for optimal efficiency) and gracefully falling back to Server-Sent Events or Long Polling as necessary. The high-level API, Hubs, facilitates seamless Remote Procedure Calls (RPC): server-side .NET code can instantly invoke client-side JavaScript functions, and vice-versa. This capability is critical for high-frequency updates in applications like live chat, real-time stock dashboards, and collaborative document editing.
Related projects
Demo: Manus — The AI Agent That Thinks Before It Acts
Cincinnati
A live demonstration of Manis, an AI agent that visualizes its reasoning steps, showing how it plans and…
Building AI react apps
Seattle
This talk demonstrates building a simple AI app using the Tambo dev tool, enabling AI to control and…
MCP Virtual Try On in C++ (KarloAIMutthiMain)
Dubai
Explore a C++ virtual try‑on app using Model Context Protocol for vector searches via Couchbase or local‑first methods,…
Artecon - A hotspot for AI
Seattle
Learn how to run CPU‑based ML models with low latency, using small public models and post‑processing, then bundle…
Invention Cards enhanced by AI
Seattle
Using GPT‑4 multimodal tools, I extracted data from Asimov’s encyclopedia, generated consistent images for a visual chronology of…
Inside “Arthur”: Wiring Multiple LLMs into a One-Click Pipeline that Generates Illustrated Novels
St Louis
This talk demonstrates how to coordinate multiple large language models and tools in a TypeScript pipeline to generate…