OpenModex
Gateway & Routing

One endpoint. Every model. The right one, every time.

An OpenAI-compatible gateway that routes each request by cost, latency, or quality — with per-source circuit breakers and automatic failover.

console.openmodex.com
OOpenModex
Overview
API Keys
Models
Routing
Provider Keys
Tools
Usage
Billing
Logs
Routing Decisions · liveConfigure
RequestStrategyRouted toLatency
chat · gpt-4o
cost_optimized
openai · direct
412ms
chat · deepseek-v3
cost_optimized
siliconflow · channel
287ms
chat · claude-sonnet
quality_first
anthropic · direct
598ms
chat · qwen-max
latency_first
novita · channel
203ms
embed · bge-m3
cost_optimized
deepinfra · failover
158ms

Three routing strategies

Choose cost_optimized, latency_first, or quality_first per request — or set a default per key.

Circuit breakers per source

Health is tracked per channel:model pair. A failing source is isolated in seconds, not minutes.

Automatic failover

When a source breaks, traffic shifts to the next healthy one mid-stream. Your users never see it.

Drop-in compatible

Existing OpenAI SDKs, LangChain, and tooling work unchanged. Change the base URL, keep everything else.

6+

providers behind one endpoint

<150ms

median added latency

3s

max time to isolate a failing source

Get up and running in minutes

Free tier included. No card required. Your first key is one click away.

Start for free