OneNexus Docs

Start with a shared Platform model, deploy a self-hosted model of your own, or connect an AI agent through Nexus Composer.

Platform models

Call models from the shared catalog through one OpenAI-compatible base URL using a Platform API key — no GPU quota required.

Self-hosted models

Deploy a model your tenant owns and operates, then call the exact endpoint and Served Model Name shown in the console.

Nexus Composer

Call OneNexus-hosted models from the agents you already use, through a local proxy.

API reference

Base URLs, authentication, request formats and error codes for the OpenAI-compatible API.