Model providers
Built-in providers
Section titled “Built-in providers”These are seeded into the database on first launch; afterwards Platform console → Model services is authoritative.
Domestic and international endpoints are separate entries — the URLs and keys are not interchangeable — so pick whichever suits your network and paste one key.
| Provider | Endpoint |
|---|---|
| DeepSeek | api.deepseek.com/v1 |
| Zhipu GLM | open.bigmodel.cn/api/paas/v4 (international: api.z.ai/api/paas/v4) |
| Qwen | dashscope.aliyuncs.com/compatible-mode/v1 (international: dashscope-intl.aliyuncs.com/compatible-mode/v1) |
| Kimi | api.moonshot.cn/v1 (international: api.moonshot.ai/v1) |
| MiniMax | api.minimaxi.com/v1 (international: api.minimax.io/v1) |
| OpenAI | api.openai.com/v1 |
| Google Gemini | generativelanguage.googleapis.com/v1beta/openai/ |
Only configured providers are shown
Section titled “Only configured providers are shown”A provider with no key never appears in a conversation’s model picker. Better than offering a provider that does nothing when you click it.
Any OpenAI-compatible service works
Section titled “Any OpenAI-compatible service works”Ollama, LM Studio, vLLM, your own gateway — if it speaks the OpenAI API, you can add it by hand in the console.
That is what makes changing providers painless: Qwen today, a local Ollama model tomorrow, one entry edited either way.
Every model call’s token usage is recorded, viewable by day, by model and by purpose, so you can keep an eye on spend.