Kition
Download

Articles · Models & providers

Configuring a custom OpenAI-compatible provider

Any endpoint following the Responses shape plugs in. Same for Ollama / vLLM / internal gateways.

Kition Team·May 1, 2026·1 min read

A generic template

{
  "name": "internal-llm",
  "baseURL": "https://llm.internal/v1",
  "apiKey": "...",
  "wire": "responses",
  "defaultModel": "internal-foundation-2026-q1"
}

Common integrations

  • Ollama: baseURL http://localhost:11434/v1
  • vLLM: ships a responses endpoint
  • Together / Fireworks etc: follow their docs
  • Internal gateway: wrap to match the responses shape
← All articles

Related reading

Models & providers

Using China-hosted models in Kition (DeepSeek / Qwen / Moonshot)

Under China network conditions, domestic models are often the most stable choice. Setup mirrors OpenAI — one baseURL away.