AI Gateway

OpenAI-compatible LLM API via LiteLLM. Local models on Arc Pro B70 and RTX 3060, with cloud fallback.

Base URL & authentication

Base URL:  https://api.kpapps.cz/ai/v1
Auth:      Authorization: Bearer

API keys are managed in LiteLLM. Contact the administrator for a key.

Available models

Model	Description	Context	Notes
`fast`	Qwen3.6 35B-A3B MTP	64K	MoE, high throughput
`smart`	Qwen3.6 27B MTP	131K	Most capable local model
`hass`	Qwen3.5 4B MTP	32K	Fast, always available
`autocomplete`	Qwen2.5-Coder 1.5B	32K	FIM only
`auto`	Cloud (OpenRouter free)	—	Cloud fallback

Quick start

# List models
curl https://api.kpapps.cz/ai/v1/models -H "Authorization: Bearer "

# Chat
curl https://api.kpapps.cz/ai/v1/chat/completions   -H "Authorization: Bearer " -H "Content-Type: application/json"   -d '{"model":"smart","messages":[{"role":"user","content":"Hello"}]}'

# Python
from openai import OpenAI
client = OpenAI(base_url="https://api.kpapps.cz/ai/v1", api_key="")