Open-source Valet runs your homelab with local LLMs across 185+ services
egfwivbwwruigwrhfbsr · reddit · 2026-11-07
A developer open-sourced Valet, a chat dashboard for managing a homelab with local LLMs — fully private, no accounts or telemetry, live demo available.
- The chat agent works with any local or hosted model, integrates with 185+ services plus custom plugins (pause pi-hole, investigate crash logs, request a movie), and can edit its own dashboard config with your approval.
- Setup: Qwen 3.8 27B on a 5090 at 230 tok/s, responses in 2–5s; passes 90% of a small eval set, but struggles with open-ended debugging and burns tool calls.
- Key optimizations (30–50% token savings): hide unused integration tools, lazy-load tool lists; disabling thinking is 2–3x faster but hallucinates follow-ups.
- Security: read-only by default, approval-gated changes, keys never exposed to the AI, Docker via read-only socket proxy.
- The author is an SWE who wrote nearly all code with Claude while owning architecture and review.
More from coding & agent
- Dev designs a cybersecurity agent that remembers past attacks with Hindsight — Adventurous-Ant-9370 · 2026-11-07
- Solo dev: most bad LLM coach answers traced to my context, not the model — Aksel_010 · 2026-11-07
- New blog: how your RL judge and credit assignment shape model reasoning quality — tokenbender · 2026-11-07
- Enzyme open-sources a personal agent memory layer saving 40% tokens vs grep — _AustinCalvert_ · 2026-11-07
- Qoder lands stealth model 'Qoduck', free daily for individual users for limited time — hey_abusiddik · 2026-11-07
- Agent autonomy should be earned in stages, not assumed, argues dev — Few-Garlic2725 · 2026-11-07