Where local .NET inference fits.
Desktop, edge, server, and regulated workloads where packaging simplicity and privacy matter.
Desktop applications
Ship an AI feature inside a WPF, WinUI, MAUI, or Avalonia app that works offline. A pure managed core keeps single-file publish and trimming intact, so the model runtime rides along with your existing installer.
Edge & embedded services
Run a 1–8B model on constrained hardware where a native toolchain, a Python sidecar, or a GPU is impractical. CPU-first execution means the same binary runs across environments.
Background workers & jobs
Summarise, classify, or extract inside a BackgroundService or a queue-driven worker. DI and Microsoft.Extensions.Hosting wiring means llmdot registers like any other service.
Privacy-sensitive & regulated work
Keep prompts and documents on the machine. Local inference means no data leaves the process boundary — useful where data residency, confidentiality, or air-gapped operation is a requirement.
ASP.NET Core integration
Expose a local model behind an endpoint via IChatClient and Microsoft.Extensions.AI. Swap a cloud chat client for a local one without rewriting your controllers or minimal APIs.
Prototyping without cloud keys
Build and demo an LLM feature with no API keys, no per-token billing, and no network dependency. Load a GGUF file and iterate locally.
Fits your workload?
Start with the quickstart, or ask us how llmdot maps to what you're building.