Lumoswitch Docs
Lumoswitch is a unified multi-model inference, control, accounting, and observability platform. Clients connect directly through enabled OpenAI, Anthropic Messages, or Gemini generateContent routes while Lumoswitch manages models, access keys, limits, routing policies, and usage records.
Start here
- Quickstart: create a configuration and make your first Lumoswitch request.
- Upstreams: add provider keys or local model servers.
- API configurations: choose a routing mode and issue an access key.
- Client setup: configure SDKs, coding agents, and web UIs.
- API reference: review the public inference routes and the Lumoswitch control API.
Core capabilities
| Capability | Description |
|---|---|
| Unified inference | Clients call one stable Lumoswitch endpoint |
| Multiple upstreams | Connect cloud and local models with provider presets or a custom OpenAI-compatible URL |
| Runtime routing | Lumoswitch executes fixed, failover, auto, least-busy, and lowest-latency configurations |
| Access and accounting | Lumoswitch enforces Access Key limits and records usage in its ledger |
New users should begin with the Lumoswitch quickstart.