Rick's crew (n8n jobs + free AI "brain switch") 🧪

The scheduled side of a home AI crew: each crew member is an n8n workflow that wakes up on a schedule, collects facts from the crew API (crew-dashboard), asks an AI to write a short in-character report, and delivers it (daily note, notification, spoken aloud).

• make_workflows.py generates the n8n workflows into workflows/ from one list of crew members (persona, schedule, model). Edit the list, run it, import with n8n import:workflow --separate --input=/workflows.
• compose.yml runs n8n (127.0.0.1 only).
• apps/compose.yml runs the free self-hosted apps the crew uses: Open WebUI, Open Notebook, Audiobookshelf, File Browser, FreshRSS, Forgejo, Stirling PDF, LibreTranslate, Kiwix (offline library), an offline map viewer and the brain switch.
• apps/brain-switch/config.yaml is a LiteLLM config: one OpenAI-style address (127.0.0.1:4000/v1) with jobs like free-chat, free-smart, free-coder. Each job tries free API tiers in order (Groq, Gemini, Mistral Codestral, OpenRouter :free, Cloudflare Workers AI, NVIDIA NIM, Ollama Cloud) and falls back to local Ollama models when offline. Put your own keys in an env file (GROQ_API_KEY=...); none are included.

Paths assume an external drive at /run/media/$USER/OmarchyExt1; change them in the compose files.

License

MIT. Built by Rabbid Raccoon with Claude. Use it, change it, share it.