Build Autonomous AI Agents That Run Entirely On-Device.
A complete agent framework for .NET: orchestrate multi-agent workflows, apply advanced reasoning strategies, connect to external services via MCP and tools, and extend capabilities with portable agent skills. All inference runs locally with full data privacy.
Multi-Agent Workflows
Pipeline, parallel, router, and supervisor orchestration patterns.
Agent Reasoning
Chain-of-thought, ReAct, tree-of-thought, and more strategies.
MCP & Tools
Full MCP client with 8 built-in tool categories and custom tools.
Agent Skills
Portable SKILL.md bundles with hot-reload and marketplace.
Compose Agents into Production Workflows.
Chain specialized agents into pipelines, fan out work in parallel, route by intent, or let a supervisor delegate and aggregate. Each agent reasons independently while the orchestrator manages handoffs and convergence.
Control How Your Agent Thinks.
Six built-in reasoning strategies let you balance speed, accuracy, and cost. From direct single-shot responses to multi-step ReAct loops with tool calls, reflection, and tree-of-thought exploration.
Connect Agents to Any External Service.
Full Model Context Protocol client implementation with 8 categories of built-in tools covering 70+ operations. Define custom tools with ITool or [LMFunction], connect to MCP servers, and compose tool chains with JSON Schema validation and parallel execution.
8 Built-in Tool Categories
Portable Skills via SKILL.md Bundles.
Package agent capabilities as portable SKILL.md files with instructions, tools, and guardrails. Load from local folders, remote URLs, or the agentskills.io marketplace with hot-reload support.
Built for Production Workloads.
Beyond orchestration and reasoning, LM-Kit.NET provides the infrastructure your agents need to operate reliably at scale.
Resilience Policies
Retry, circuit breaker, rate limiting, timeout, fallback, and bulkhead policies to keep agents running in adverse conditions.
Observability & Tracing
Full agent tracing with spans, metrics, and JSON export. Console, in-memory, and composite tracers for debugging and monitoring.
Real-Time Streaming
Buffered, delegate, multicast, and text writer stream handlers for token-by-token agent output with full async support.
16 Agent Templates
Pre-built templates for assistant, analyst, researcher, reviewer, code, debugger, editor, extractor, QA, tutor, and more.
Agent Memory
RAG-powered persistent memory that lets agents recall context across sessions using vector search and embeddings.
Agent Delegation
Name-based routing lets a lead agent dynamically delegate tasks to specialist sub-agents with full result aggregation.
The Most Complete Agent Framework for .NET
From single-agent prototypes to complex multi-agent pipelines with resilience, tracing, and streaming. LM-Kit.NET gives you everything you need to ship production agents that run entirely on-device.