VI1’s own API
OpenAI-compatible endpoints backed by GLM, with VI1’s exact calculator, private knowledge retrieval and JSON answer checking built in. Keys, rate limits, a hard monthly spend cap and response caching keep usage under control.
Drop-in compatible
POST /v1/chat/completions with streaming, GET /v1/models. Works with any OpenAI SDK: set base_url to this server.
Provider switch
Z.ai GLM today; your own vLLM/SGLang GLM server later. One config line, with automatic fallback.
VI1 modules
Exact calculator, knowledge retrieval, checked calculations and schema-validated JSON, visible in every response.