
Run powerful large-language models entirely on your player’s machine—no servers, no cloud fees, no compliance headaches.If you are running on some older CPU SIMD instruction sets you might encounter some issues. We have a fix for this on our github and are working on creating a way to get this to the asset store as well.NobodyWho wraps the battle-tested llama.cpp runtime in a clean Unity-ready API, giving you:Drop-in C# bindings – call chat, completion, or embedding functions from any script or add them as components directly in you scene.GPU-accelerated inference – Vulkan (Windows/Linux) & Metal (macOS) back-ends for blazing-fast token generation.Zero MLOps – every token is computed client-side, so you ship once and scale automatically.Tiny footprint – lean native library - you bring your own model.Open-source core – inspect, fork, and extend on GitHub.Get started in minutes with the quick-start sample scene, or dive deeper in our docs.Focus on the experience you’re building—NobodyWho handles the LLM heavy lifting.Please help us by giving us a star on our GithubKey Features:🏃 Offline CharactersGenerate dialogue, decisions, and narration entirely on-device.⚡ Blazing-Fast GPU InferenceVulkan (Windows / Linux) and Metal (macOS) back-ends, with CPU fallback on older hardware.🔧 Two-Object SetupDrop NobodyWhoModel and NobodyWhoChat into your scene, assign a model file and send a message - that is it.🎯 Built for Interactive ContentPerfect for NPC brains, branching stories, procedural quests, or chat-driven interaction.💻 Cross-PlatformOne native library covers Windows, Linux, and macOS; build once, ship everywhere.⏩ Streaming TokensWords appear in real time📦 Tiny Footprint~20 MB native library—just add the model files you need.👩💻 Open-Source & PermissiveEUPL-licensed —fork, extend, or ship commercially without extra fees.🛠️ Sample Scene & DocsQuick-start chat UI, code snippets, and a step-by-step install guide included.🦙 Powered by llama.cppSupports GGUF compatible models