The Private Document Intelligence Platform

From documents to data, knowledge, and action. Privately.

LM-Kit turns documents into structured data, internal knowledge into cited answers, and document workflows into governed automation, entirely on infrastructure you control. No cloud calls, no data leaving your machines, no per-token fees.

---

🧩 Two Products, One Engine

🖥️ LM-Kit One - A private AI server you deploy. Structured extraction, knowledge search with citations, governed document agents and model management, reached over native REST, OpenAI-compatible, Ollama and Anthropic APIs, or from AI assistants over MCP.

📦 LM-Kit.NET - The Private Document Intelligence SDK for .NET. The same engine as one NuGet, fully in-process: structured extraction with per-field confidence, OCR, PII redaction, RAG with citations, agents, vision and speech. Zero cloud calls.

---

🤗 Models in This Organization

These repositories are the LM-Kit model catalog: curated, quantized builds maintained for both products.

• *-gguf - Standard GGUF quantizations, usable with any GGUF-compatible runtime
• *-lmk - The same models packaged in the LM-Kit format, loaded directly by model ID, for example LM.LoadFromModelID("qwen3.8:27b")

---

✨ What You Can Build

📄 Documents to data - Schema-driven extraction with per-field confidence, OCR, layout analysis, Markdown conversion
🔍 Knowledge with citations - Hybrid search and RAG over internal documents, with answers citing document and page
🛡️ Privacy workflows - PII detection, redaction, classification, named entity recognition
🤖 Governed agents - Tool calling under permission policies, memory, MCP
🎨 Vision - Visual document understanding, image embeddings, multimodal search
🎙️ Speech - Local speech-to-text and transcription

---

🔒 Why Private?

• Private by architecture - Inference runs on your hardware; offline and air-gapped deployments work
• Verifiable by output - Extractions carry confidence scores, answers carry citations
• Governed in operation - Permission policies decide what agents and tools may do
• Cross-platform - Windows, Linux, macOS (including ARM64)
• No per-token fees - Predictable licensing, unlimited inference

---

📚 Resources

🌐 Website: lm-kit.com
📖 Documentation: docs.lm-kit.com
💻 GitHub Samples: LM-Kit Samples
📦 NuGet: LM-Kit.NET Packages
🛡️ Trust Center: lm-kit.com/trust

---

🎁 Free for Small Businesses and Education

Our mission is to empower every organization with private AI. Both products are free for small businesses and for education, including commercial use, with nothing to activate and no runtime license check. Evaluation and development are free at any company size.

👉 Free Download | LM-Kit One | Pricing