chalard.dev

ฉลาด — a small self-hosted LLM lab. Local models on an old Xeon, remote models when it matters. Private; sign-in required.

chat
Open WebUI →
Talk to local Gemma / Qwen or GPT · Claude · Gemini through one UI.
observe
Grafana →
Tokens/s, model load state, host temps and power.
build
API →
OpenAI-compatible endpoint via LiteLLM. Bring your own key.

quick start

# OpenAI-compatible · base_url = https://api.chalard.dev/v1
curl https://api.chalard.dev/v1/chat/completions \
  -H "Authorization: Bearer $CHALARD_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gemma4-e4b","messages":[{"role":"user","content":"สวัสดี"}]}'

# models: gemma4-e4b · qwen3-30b-a3b · gpt-5-mini · claude-sonnet · gemini-flash · or-free