Homelab RAG — Private AI Q&A for your self-hosted docs
# Homelab RAG — Private AI Q&A for your self-hosted docs ### Stop grepping your wiki. Just ask it. You've documented your homelab — Jellyfin quirks, backup runbooks, that Postgres restore you figured out at 2am. Six months later you can't find any of it. **Homelab RAG** turns your **Outline / Paperless** docs into a private AI assistant you ask from Telegram. Accurate answers, **with the source doc linked**, running **on your own server**. > "How do I set up Jellyfin hardware transcoding?" → the exact steps from **your** runbook, source linked. **▶️ Watch the 2-min demo:** https://www.loom.com/share/e2dc707390d742bcabda154af684eb7f --- ### Why not just AnythingLLM / Open WebUI / ChatGPT? - ✅ **Auto-syncs** your docs as they change (they don't) - ✅ Pulls straight from **Outline / Paperless** (they don't) - ✅ **Multilingual** embeddings — works on non-English docs (most fail here) - ✅ **Always cites** the source document - ✅ **No hallucination** — says "Not found" instead of guessing - ✅ Yours to extend — it's **n8n** ### What you get - 2 ready-to-import **n8n workflows** (auto-indexer + Telegram Q&A agent) - `docker-compose.yml` — full stack (n8n + Ollama + Qdrant) - **Local mode** (private) *and* **cloud mode** (fast) — your choice, per query - Step-by-step SETUP with **honest** performance numbers - Outline + Paperless connectors; extend to any API ### Honest about speed Local answers on a CPU box take ~60–90s (privacy has a cost). Want instant? Use a GPU, or flip to cloud mode (~3–5s) — **retrieval always stays local**, only the question + excerpts go out. Your call. ### Who it's for Self-hosters running Outline/Paperless who want their knowledge back — without handing it to a cloud. --- **Launch price: $39** (pre-order — you get every update). 30-day refund if it doesn't work for you.
Get it → napassadol.gumroad.com