NexusAPI — Unified AI Gateway & Model Router
NexusAPI is a self-hosted AI gateway that unifies more than forty upstream large-language-model
providers — including OpenAI, Anthropic Claude, Google Gemini, AWS Bedrock, Azure OpenAI,
Mistral, Cohere, DeepSeek, Qwen and others — behind a single OpenAI-compatible REST API. It lets
product teams point their existing OpenAI SDK calls at one endpoint and route traffic to any
supported upstream provider, with built-in intelligent routing, automatic retry on upstream
failures, channel weighted load balancing, per-token usage metering, group-based billing ratios,
and a built-in admin dashboard for users, channels, tokens, redemptions, logs and quota.
Key capabilities
- 40+ upstream providers behind a single OpenAI-compatible API, including Chat Completions, Responses, Realtime audio, Embeddings, Rerank, Images, Audio and Midjourney/Suno relays.
- Format conversion between OpenAI-compatible, Anthropic Messages and Google Gemini formats, so clients written for one vendor can call any other.
- Intelligent routing with channel weighting, automatic failure retry, per-user and per-model rate limits, and tiered billing expressions for dynamic pricing.
- Built-in metering & billing with token-level quota tracking, group ratios, cache-token billing, online recharge (EPay, Stripe) and an expression-based pricing engine.
- Authentication & SSO via JWT, WebAuthn / Passkeys, GitHub, Discord, LinuxDO, Telegram, and OpenID Connect, plus per-token group quota.
- Admin dashboard with visual statistics, channel management, user management, log audit and a multi-language user interface.
- Self-hostable with SQLite, MySQL (≥ 5.7.8) and PostgreSQL (≥ 9.6) backends, Redis cache, multi-machine deployment and Docker / docker-compose images.
OpenAI-compatible endpoints
Clients using the OpenAI SDK can call /v1/chat/completions, /v1/responses,
/v1/embeddings, /v1/images/generations, /v1/audio/*,
/v1/rerank and /v1/realtime without code changes; Claude-format clients can
call /v1/messages and Gemini-format clients can call the Gemini relay root. The gateway
normalises authentication, request shape, streaming, tool use and error handling across all upstreams.
Getting started
Deploy NexusAPI with docker run or docker-compose up -d, then sign in with
the root account and add your upstream provider credentials as channels. Create API tokens for your
applications, attach a billing group, and point your existing OpenAI SDK call at
https://your-host/v1. The dashboard at the root path lets you monitor usage, manage
channels and review logs in real time. A full deployment guide, environment-variable reference and
API reference are available in the official documentation.
This page is the admin dashboard for the NexusAPI gateway. To explore the public, ad-supported
free online tools hosted on adjacent sub-domains, visit
tools.aiprh.com.