Runnable C# examples for LM-Kit.NET, the local-first AI SDK for .NET.
Every sample runs offline, on your own hardware, with no API keys and no cloud calls. Clone, dotnet run, done.
New to LM-Kit? Start at lm-kit.com for the product overview, or jump straight into a sample below.
Samples are organized by LM-Kit capability pillar. Open any folder for runnable projects and per-sample readmes.
| Pillar | What you'll find |
|---|---|
| AI Agents | Agent orchestration, planning strategies, tools, function calling, MCP, multi-agent workflows, memory, skills |
| Document Intelligence | PDF chat, classification, splitting, summarization, structured field extraction, document-to-markdown |
| Vision & Multimodal | Image chat, image embeddings, document OCR (traditional and VLM-based), document layout extraction |
| RAG & Knowledge | RAG chat, vector databases, data extraction pipelines |
| Text Analysis | Classification, NER, PII, sentiment, keyword extraction, language detection |
| Speech & Audio | Real-time and batch speech-to-text |
| Text Generation | Conversations, translation, summarization, prompt templates, grammar correction |
| Local Inference | Encrypted models, runtime configuration |
| Model Optimization | Quantization, fine-tuning |
| Integrations | Microsoft.Extensions.AI, Semantic Kernel |
git clone https://github.com/LM-Kit/lm-kit-net-samples.git
cd lm-kit-net-samples/console_net/<pillar>/<sub-category>/<sample>
dotnet runEach sample:
- Targets .NET 8 or later
- Pulls the LM-Kit.NET NuGet package automatically
- Downloads its model on first run (cached for later runs)
- Picks the fastest backend available on your machine (CUDA, Vulkan, Metal, or CPU)
No license key, no signup, no configuration. When you're ready to build your own app, grab a free Community Edition license.
LM-Kit auto-selects the best backend at startup. You don't configure anything.
- NVIDIA, AMD, Intel GPUs
- Apple Silicon (Metal)
- Multi-GPU and hybrid CPU+GPU for models larger than VRAM
- CPU-only fallback with AVX/AVX2
Sample readmes list VRAM expectations where it matters.
© LM-Kit