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.
Read more →Per-task model selection: balancing cost and quality
Not every task needs opus / GPT-5. Give light tasks haiku / mini; reserve opus for hard ones.
Read more →Model usage dashboard: cost control in practice
Break usage down by provider / model / vault — spot AI-field black holes and long-context abuse.
Read more →Configuring a custom OpenAI-compatible provider
Any endpoint following the Responses shape plugs in. Same for Ollama / vLLM / internal gateways.
Read more →What is AiApiy and how to set it up in Kition
AiApiy is a unified reverse-proxy across OpenAI / Claude / Gemini. One account, many providers.
Read more →Driving the Kition agent with GPT-5
GPT-5 is rock-solid on structured outputs and multi-step function calls. A great fit for OpenAI-native workflows.
Read more →Driving the Kition agent with Claude 4.7
Claude is consistently strong on long context, tool use, and alignment. A solid default for the Kition agent.
Read more →OpenAI Responses vs Anthropic Messages: how to pick
Both wires are mature. The differences live in tool-call patterns, streaming preference, and pricing band.
Read more →Ready when you are.
Kition is a local-first AI workspace. Markdown documents, structured tables, and an AI agent — running on your own machine, against the model provider you choose.