ODS — the Osmantic Deployment System — turns an ordinary PC, Mac or Linux box into a private AI server. One installer sets up local model inference, a chat interface, voice, agents, RAG and image generation as a single wired-together stack, so you do not have to assemble the pieces by hand. The project has about 7,600 stars on GitHub, is released under the Apache 2.0 license, and its V3 generation is in active pre-release development in 2026.
ODS installs and configures the components of a home AI stack for you: llama.cpp-based local inference, Open WebUI for ChatGPT-style conversation, n8n for workflow automation, ComfyUI for image generation, plus speech, search and retrieval services — all managed from one control dashboard. The installer auto-detects your hardware, from NVIDIA GPUs and AMD Strix Halo to Apple Silicon and Intel Arc, and picks suitable models and settings. Privacy is the design center: no cloud is required, no telemetry is collected, and chats, files and inference stay on your machine, while optional cloud or hybrid modes can call OpenAI, Anthropic or Together APIs when you prefer. A guided Windows installer (via WSL2) sits alongside the one-command Linux and macOS setup.
Free. ODS is open source under the Apache 2.0 license and explicitly requires no subscriptions; any cost is your own hardware and, optionally, third-party API usage.
Homelab builders, privacy-conscious users and small teams who want the capabilities of a cloud AI subscription running entirely on hardware they own, without days of manual configuration.