cairn
Cairn is a self-hosted wellness OS that reads labs, lifting, running, food, sleep and life into one agentic brain suggesting the one thing worth doing today. It runs on your own hardware with SQLite storage, no scores or nudges.
Ficha actualizada 22 sept 2026, 20:52 · upstream 22 sept 2026, 20:49 · release v1.8.0
Qué es cairn
Descripción editorial construida a partir del repositorio, su documentación y sus señales de mantenimiento.
Cairn is a self-hosted health coach that treats your body as an ecosystem rather than a collection of metrics. The app reads your lab results, lifting sessions, running data, food intake and sleep patterns into one coherent picture. It then suggests the single most important thing to do today — not because it's nudging you, but because it understands what matters to you.
The architecture is agent-native: a full MCP server runs alongside the PWA so any client can read or write data. The app never scores you, never tells you what to eat, and never gives advice without context. It works by reading your history, finding patterns in plain language, and presenting one actionable suggestion at a time.
The design philosophy is simple: calm over control. You set goals and constraints, the system does the heavy lifting of synthesis, but you always drive the direction. This makes it suitable for people who want to understand their own habits rather than be told what they should do.
- A user with a new lab result (e.g., elevated cholesterol) wants to know exactly how to adjust their diet and training without feeling judged or overwhelmed by the numbers
- An athlete who is confused about whether their current recovery strategy is working needs a clear, plain-language explanation of what's happening in their body
- A person wanting to improve long-term health goals (like weight loss or endurance) wants concrete steps that don't rely on abstract 'score' targets but instead build toward real outcomes
- Someone who struggles with motivation and feels like they're not making progress needs a system that frames actions around their own values rather than external metrics
- A user wanting to understand the relationship between sleep, recovery, and performance wants a visual representation of how different variables interact over time
- An individual seeking to optimize both strength and longevity wants a unified approach that considers all aspects of health simultaneously without conflicting advice
- Cairn runs entirely on your own hardware with data stored locally in SQLite, meaning no cloud dependency or third-party data collection is required
- The app uses an agent-native architecture where a full MCP server communicates with the PWA, allowing any client to read and write data without additional setup
- All suggestions are generated from your actual historical data rather than arbitrary targets, making them personally relevant and less likely to feel like generic advice
- The app requires a minimum of Node.js version 24.0 or higher for the sqlite driver to function correctly
- While the Brief provides general suggestions, it does not provide personalized medical advice or diagnose specific health conditions
- The app does not integrate with third-party fitness platforms like Apple Health or Google Fit, limiting data sources beyond what is manually logged
- The agent-based features (chat and meal planning) require manual configuration of an agent provider and do not work out-of-the-box for all use cases
- The system relies on user input for lab results and training details, meaning it cannot autonomously collect or interpret data without explicit user action
Perfil de infraestructura
Estimación inicial derivada del repositorio y de los servicios detectados.
Despliegue detectado
Qué señales reales de instalación hemos encontrado.
Cairn is deployed using Docker Compose with a single Node.js service that serves the PWA, REST API, MCP server, and background scheduler. The app uses SQLite via Node's built-in sqlite3 module for all data storage. For production deployment, consider running on a dedicated server or VM rather than a shared host to prevent other services from interfering with health-related requests. Set up environment variables for your timezone (e.g., TZ=America/New_York) and configure the MCP server with appropriate authentication if you plan to use it externally. The container image is multi-arch, so ensure your deployment infrastructure supports amd64 and arm64 architectures.
Stack y dependencias
Componentes inferidos a partir de manifests, configuración y documentación.
Puede sustituir a
Relaciones detectadas por el análisis del repositorio y su categoría.
Experiencia real de la comunidad
Ayuda a contrastar la estimación automática con instalaciones reales.
Todavía no hay datos comunitarios para este proyecto.