muneeb__mughal 

A batteries-included local LLM inference library for .NET. Drop-in Microsoft.Extensions.AI IChatClient backed by LLamaSharp — run GGUF models locally with a single AddHearth() call.

ASP.NET Core integration for Hearth — OpenAI-compatible /v1 endpoints (chat completions, embeddings, models) via a single MapHearth() call.

Apple Metal GPU backend for Hearth. Install this package instead of the base Hearth package to enable GPU-accelerated inference on Apple Silicon (M1/M2/M3/M4) and Intel Macs with AMD GPUs.

NVIDIA CUDA 12 GPU backend for Hearth. Install this package instead of the base Hearth package to enable GPU-accelerated inference on CUDA 12-capable NVIDIA hardware.

Vulkan GPU backend for Hearth. Install this package instead of the base Hearth package to enable GPU-accelerated inference on AMD, Intel, and other Vulkan-capable GPUs.

RAG (retrieval-augmented generation) pipeline for Hearth.AI — index documents, search by embedding similarity, and generate grounded answers from local LLMs.

dotnet new templates for Hearth.AI — local LLM inference for .NET. Includes a Web API template that exposes OpenAI-compatible endpoints backed by a local GGUF model.

Embeddable streaming chat UI component for Blazor Server and WASM — powered by Hearth.AI local LLMs.

Aspire hosting integration for Hearth.AI — adds a Hearth inference server as an Aspire resource with one AddHearth() call.

Aspire client integration for Hearth.AI — configures IChatClient from an Aspire-injected connection string pointing at a Hearth inference server.