AI Agent News/stories/edi-life-os-self-hosted-life-dashboard-with-an-mcp-server-for-ai-assistants.md

MCP

Edi Life OS: self-hosted life dashboard with an MCP server for AI assistants

Edi Life OS is an open-source, self-hosted dashboard that combines habits, goals, Kanban boards, finance, focus timers and notes in one PHP and MySQL app. It includes an optional MCP server that lets assistants such as Claude read the dashboard and make changes through the app's token-protected HTTP API.

Source: github.com/edrisranjbar/lifeos

The project runs on any PHP and MySQL host, including shared hosting, and can be started with Docker Compose. Its ten workspaces share one sign-in and one database, and the Overview combines focus, habits, goals and work into a daily score.

The MCP server exposes 15 tools covering the dashboard, goals, tasks, habits, finances and notes. It runs over stdio, and according to the README it talks to the app only through the HTTP API, never touching the database or credentials directly. Setup uses LIFEOS_BASE_URL and LIFEOS_API_TOKEN, and the README lists Claude Desktop, Claude Code and Codex as supported clients.

The README says the API token grants read and write access to every supported domain, and that token scopes and app-level rate limiting are not implemented. Request limits therefore need to be set at the hosting layer, and the project is built for a single owner. The code is MIT licensed and the repository showed 68 stars at the time of capture.

Why it matters if you build agents

Developers who want a personal data assistant can point an MCP client at a self-hosted instance and have it log habits or expenses without a hosted third-party service.

Because the API token has full read and write access and has no scopes, anyone wiring this into an agent should treat the token as a full-owner credential and put rate limits and HTTPS in front of it.

Get stories like this every weekday

One short email with the day's agent tooling news.

Subscribe