# SciSharp/LLamaSharp

A C#/.NET library to run LLM (🦙LLaMA/LLaVA) on your local device efficiently.

Repository: https://github.com/SciSharp/LLamaSharp
Canonical: https://ross.abutalabs.com/products/llamasharp
Homepage: https://scisharp.github.io/LLamaSharp
Language: C#
License: MIT
License Family: permissive
Topics: chatbot, gpt, llama, llamacpp, llm, semantic-kernel, llava, multi-modal, llama2, llama3, llama-cpp
Last push: 2026-08-24T19:02:52+00:00

## Health v2 (maintenance only)
Score: 92/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 87, longevity 86
- inputs: {"age_days": 1212, "days_push": 9, "days_rel": 9, "gap_med": 62.5, "n_releases_24m": 11}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3787, forks 506 (observed 2026-08-28T04:08:18.651616+00:00)

## What it is
LLamaSharp is a cross-platform C#/.NET library that wraps llama.cpp to run LLaMA-family and other LLMs (including multimodal LLaVA) efficiently on local devices with CPU, CUDA, or Vulkan backends. It provides high-level APIs, chat sessions, and RAG support for deploying LLMs in .NET applications.

## Use cases
- run llama models locally in a C# app
- run LLM inference on device with .NET
- build a chatbot in C# with a local LLM
- run LLaVA multimodal model in .NET
- integrate local LLM with Semantic Kernel
- do RAG with a local LLM in .NET
- run LLM on GPU with CUDA from C#

## When to choose
- you are building a .NET/C# application that needs local LLM inference
- you want llama.cpp performance without writing C++ or managing native bindings yourself
- you need CPU, CUDA, or Vulkan backend flexibility on cross-platform targets
- you want high-level chat, RAG, and Semantic Kernel integration out of the box

## When to avoid
- your stack is Python, Node.js, or another non-.NET ecosystem
- you need to serve many concurrent users at scale rather than local/on-device inference
- you need models or features not supported by llama.cpp
- you want a managed pure-C# inference engine without native binaries

## Facets
- artifact type: library
- maturity: active
- function: llm-inference, rag, chatbot, sdk
- domain: large-language-models, artificial-intelligence, machine-learning, chatbots
- platform: cross-platform, windows
- tags: llamacpp, dotnet, csharp, local-inference, llava, multimodal, semantic-kernel, retrieval-augmented-generation, linux, macos, gpu

## Member repositories
- SciSharp/LLamaSharp (main) score 92

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:18.651616+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:27:28.339452+00:00, confidence not recorded.
  - readme: https://github.com/SciSharp/LLamaSharp (fetched 2026-08-28T04:08:18.651616+00:00, sha 13fd53d083e4)
  - homepage: https://scisharp.github.io/LLamaSharp (fetched 2026-08-29T09:22:06.011105+00:00, sha 0c2cd705c1ad)
- Data as of 2026-08-30T08:39:29.467469+00:00.