Everything LM-Kit ships, in one place.
LM-Kit is a single NuGet package that covers a broad surface of generative-AI capabilities, every category below is a feature already available in the SDK. Pick the area you need, the rest stays out of your way.
01 · Agents
AI agents on local models
Planners, supervisors, parallel and pipeline orchestrators, persistent memory, MCP clients, custom tool calling, agent skills.
AI Agents02 · Documents
Document intelligence
PDF parsing, OCR, structured field extraction, chat-with-PDF, document RAG, splitting.
Document Intelligence03 · RAG
Retrieval & knowledge
Built-in vector store, Qdrant and pgvector connectors, hybrid retrieval, RAG chat, data extraction.
RAG & Knowledge04 · Embeddings
Dense & sparse vectors
Qwen3-embedding, EmbeddingGemma, dense and sparse retrieval, on-device.
Embeddings05 · NLP
Text analysis & classification
Classification, NER, PII extraction, sentiment, keyword extraction, language ID.
Content analysis06 · Generation
Translation, rewriting & more
Translation, summarisation, correction, rewriting, schema-constrained generation.
Text generation07 · Speech
Audio transcription, STT
A growing local speech-to-text stack: on-device, multilingual, streaming output, real-time translation, 100+ languages.
Speech to text08 · Optimization
Quantization & fine-tuning
Quantize open-weight models, hot-swap LoRA adapters, deterministic and grammar-constrained sampling. Make your model fit your hardware.
Model optimizationGet started