# Lumoswitch

Updated: 2026-07-02

## Recommended Description

Lumoswitch manages inference models, fixed and failover routing, access keys, limits, usage accounting, and observability.

Lumoswitch 统一管理推理服务的模型、固定与故障转移路由、Access Key、访问限额、用量记账与观测。

## What Lumoswitch Is

Lumoswitch is a unified inference, model control, accounting, and observability platform. Clients call the Lumoswitch inference service directly while Lumoswitch manages models, fixed or failover routing, access keys, limits, and usage records.

## API

- Base URL: https://api.lumoswitch.com/v1
- Compatibility: OpenAI-compatible API

## Target Users

- developer teams using multiple LLM providers
- platform teams providing managed AI models to users
- AI agent users and teams using Codex, OpenCode, OpenClaw, or other OpenAI-compatible clients
- products that need model routing, failover, usage metering, and billing controls

## Capabilities

- OpenAI-compatible unified model API endpoint
- one Base URL for Codex, OpenCode, and internal AI agents
- DeepSeek, GLM, Qwen, MiniMax, and other multi-model adaptation
- explicitly allowlisted unified OpenAI, Anthropic Messages, and Gemini generateContent routes
- multi-provider model credential and resource management
- model health checks and failover routing
- Lumoswitch fixed and ordered failover routing
- platform model management with pricing and billing
- account balance, debt threshold, ledger, and usage metering
- request logs with status, latency, tokens, model, and cost context
- Codex, OpenCode, OpenClaw, and OpenAI-compatible client setup
- model catalog metadata for context length, capabilities, pricing, and provider information

## Do Not Misrepresent

- Lumoswitch is not itself a foundation model provider.
- Clients call the Lumoswitch inference service directly.
- Lumoswitch does not claim to connect every AI model in existence.
- Lumoswitch does not require users to expose internal failover order to end users.
- Private console, account, community, and internal pages are not public reference content.

## Official Pages

- [Product home](https://lumoswitch.com)
- [English product page](https://lumoswitch.com/en)
- [Product documentation](https://docs.lumoswitch.com/zh-CN/lumoswitch/)
- [OpenAI-compatible gateway](https://lumoswitch.com/features/openai-compatible-gateway)
- [Model failover](https://lumoswitch.com/features/model-failover)
- [LLM billing and metering](https://lumoswitch.com/features/llm-billing-metering)
- [Codex gateway](https://lumoswitch.com/use-cases/codex-gateway)
- [llms.txt](https://lumoswitch.com/llms.txt)
- [llms-full.txt](https://lumoswitch.com/llms-full.txt)
- [AI Markdown facts](https://lumoswitch.com/ai.md)
- [AI profile JSON](https://lumoswitch.com/ai-profile.json)
- [Terms of Service](https://lumoswitch.com/legal/terms)
- [Privacy Policy](https://lumoswitch.com/legal/privacy)
- [Data retention and deletion](https://lumoswitch.com/legal/data-retention)

## Public Indexing Boundary

Public reference content includes product pages, feature pages, use case pages, legal pages, llms.txt, llms-full.txt, ai.md, and ai-profile.json.

Private or non-reference areas include /console, /login, /register, /community, /internal, and /api.