Beelink Agent Stack Kit — Run a Local AI Agent System on a $189 Mini PC

This is the DAGR architecture I built and run daily on a Beelink Mini PC. Seven months of iteration, packaged into a drop-in kit. It runs n8n workflows, a local LLM via Ollama, systemd timers, a Telegram bot, and a GOAL.md context framework that gives your agent behavioral rules it reads on every fire. The whole system runs 24/7 on 8W of power, zero API costs, zero cloud dependency.

What you get: every n8n workflow JSON, directives.py with inline comments, the full GOAL.md framework template, systemd timer configs, Telegram bot setup guide, and a step-by-step install doc that assumes you know how to SSH but nothing more. If you want a real agent system running on your own hardware — not a demo, not a tutorial, a live working stack — this is it. Price: $97, one-time.

What's included:

• Full DAGR architecture: n8n workflow JSONs, directives.py, GOAL.md framework, systemd timers, and Telegram bot config all in one download.
• - Zero ongoing API cost: runs Mistral 7B or Llama 3 locally via Ollama. The Beelink paid for itself in under three weeks of avoided API spend.
• - Built from 7 months of live iteration: every file in this kit is from a system running 24/7 with 100%+ uptime logs, not a tutorial project.
• - Telegram delivery layer included: structured digests land in your phone before you touch it. Morning briefings, DAGR loop results, health data summaries, all routed through a Telegram bot.
• - Designed for solo builders: SSH access assumed, nothing else. No DevOps background needed. Install doc takes you from fresh Ubuntu to first DAGR fire in under two hours.