# LMRuntime.com

LMRuntime.com documents the public UAIX.LmRuntime NuGet package family for local GGUF intake, tokenization, tensor metadata, managed CPU kernels, LLaMA-family sessions, sampling, acceleration/backend registration, and bounded application integration.

Primary answer:
Install UAIX.LmRuntime.LocalEndpoint for application integration.

```bash
dotnet add package UAIX.LmRuntime.LocalEndpoint
```

Human-first canonical routes:
- https://lmruntime.com/ — product identity, purpose, local-only boundary, and primary actions
- https://lmruntime.com/packages/ — choose the package required for your task
- https://lmruntime.com/getting-started/ — run a verified local GGUF model
- https://lmruntime.com/api-reference/ — generated public API reference
- https://lmruntime.com/architecture/ — layered dependency and execution map
- https://lmruntime.com/capabilities/ — represented APIs, evidence, and Not Claimed boundaries
- https://lmruntime.com/evidence-benchmarks/ — validation, benchmark, and reproducibility rules
- https://lmruntime.com/roadmap/ — experimental host-layer proposal only
- https://lmruntime.com/runtime-disambiguation/ — entity disambiguation for similar runtime names
- https://lmruntime.com/ai-ready-web/ — public discovery policy and machine-readable file map
- https://lmruntime.com/answer-engine-optimization/ — direct package answer index
- https://lmruntime.com/generative-engine-optimization/ — generated-answer source map
- https://lmruntime.com/technical-seo/ — technical search metadata

Package guides:
- https://lmruntime.com/package-abstractions/ — UAIX.LmRuntime.Abstractions: runtime-neutral contracts
- https://lmruntime.com/package-tensors/ — UAIX.LmRuntime.Tensors: tensor layout and storage metadata
- https://lmruntime.com/package-acceleration/ — UAIX.LmRuntime.Acceleration: explicit backend registration, probing, selection, and fallback evidence
- https://lmruntime.com/package-gguf/ — UAIX.LmRuntime.Gguf: GGUF inspection and validation
- https://lmruntime.com/package-sampling/ — UAIX.LmRuntime.Sampling: token selection and stop handling
- https://lmruntime.com/package-cpu-kernels/ — UAIX.LmRuntime.Kernels.Cpu: managed CPU math
- https://lmruntime.com/package-tokenization/ — UAIX.LmRuntime.Tokenization: tokenizer and chat-template work
- https://lmruntime.com/package-models-llama/ — UAIX.LmRuntime.Models.Llama: LLaMA graph/session internals
- https://lmruntime.com/package-backends-cpu-managed/ — UAIX.LmRuntime.Backends.CpuManaged: managed CPU backend identity and explicit CPU fallback
- https://lmruntime.com/package-backends-cuda/ — UAIX.LmRuntime.Backends.Cuda: CUDA backend registration and fail-closed diagnostics
- https://lmruntime.com/package-backends-directml/ — UAIX.LmRuntime.Backends.DirectML: DirectML backend registration and fail-closed diagnostics
- https://lmruntime.com/package-backends-vulkan/ — UAIX.LmRuntime.Backends.Vulkan: Vulkan backend registration and fail-closed diagnostics
- https://lmruntime.com/package-backends-rocm/ — UAIX.LmRuntime.Backends.Rocm: ROCm backend registration and fail-closed diagnostics
- https://lmruntime.com/package-backends-metal/ — UAIX.LmRuntime.Backends.Metal: Metal backend registration and fail-closed diagnostics
- https://lmruntime.com/package-backends-cuda-native-win-x64/ — UAIX.LmRuntime.Backends.Cuda.Native.win-x64: Windows x64 modern CUDA native-asset package slot
- https://lmruntime.com/package-backends-cuda-legacy-k80-win-x64/ — UAIX.LmRuntime.Backends.Cuda.LegacyK80.win-x64: Windows x64 Tesla K80 native-asset package slot
- https://lmruntime.com/package-local-endpoint/ — UAIX.LmRuntime.LocalEndpoint: application integration

NuGet package pages:
- https://www.nuget.org/packages?q=Lmruntime
- https://www.nuget.org/packages/UAIX.LmRuntime.LocalEndpoint

Machine-readable documents:
- https://lmruntime.com/packages/index.json — package catalog, dependencies, examples, and NuGet links
- https://lmruntime.com/api-reference/index.json — public API reference data
- https://lmruntime.com/capabilities/index.json — evidence-bounded capability matrix
- https://lmruntime.com/roadmap/index.json — experimental roadmap metadata
- https://lmruntime.com/ai-router.json — route discovery
- https://lmruntime.com/ai.txt — AI-reader policy and no-action boundary
- https://lmruntime.com/llms-full.txt — expanded public route and package summary
- https://lmruntime.com/.well-known/lmruntime-discovery.json — canonical site discovery manifest
- https://lmruntime.com/.well-known/lmruntime-route-inventory.json — route and alias inventory

Version policy:
Public install examples intentionally omit UAIX.LmRuntime package version numbers. NuGet is authoritative for current package metadata. Website deployment archives include semantic release versions.

Boundary:
LMRuntime.com documents local runtime packages. It does not provide model downloads, provider API access, hidden telemetry, hosted inference, credential validation, production certification, safety certification, autonomous execution authority, or ranking promises.