https://routing.dari.dev/v1 with a Routing API key and model: "dari/routing". The /v1 path uses your organization’s default router, which starts with Dari’s managed model set and requires no setup.
Choosing A Model
dari/routing lets the router choose a model and thinking level for each request. Choosing between more than one model or thinking level needs the Autorouting plan; without it, those requests return 402 autorouting_plan_required.
To use one model yourself, send model: "dari/model/<model_id>" with any model enabled on the router, for example dari/model/anthropic/claude-opus-5-5. The request is served by exactly that model, without selection, on any plan. The thinking level you request is used when the model supports it; otherwise the model’s default applies. A model the router does not enable returns 400 model_not_enabled. Codex lists these ids in its /model menu.
Address a specific router by its rtr_... ID instead of /v1. Both forms accept the same request bodies, and endpoint paths vary by protocol:
Claude Code, Codex, OpenCode, Pi, OpenClaw, and Hermes Agent take base URLs and append their protocol-specific paths, so use the base URL shown in each coding-agent guide.
Chat Completions API
OpenAI-compatible Chat Completions.
Responses API
OpenAI Responses over HTTP.