Manifesto
Why we route.
The story of routed.sh and why AI infrastructure needs a rethink.
The Problem
Every developer using AI inference hits the same wall. You subscribe to a coding AI plan. You get a monthly request limit. You get the models that provider chose for you. And when their infrastructure has issues — which it does — you sit there waiting. Your editor is slow, your completions hang, your agent stalls. You can't do anything about it because you're locked into one provider.
So you consider adding another subscription. But now you're paying two monthly fees, managing two accounts, and still writing your own fallback logic. Most developers just accept the downtime and overpay for the uptime.
Meanwhile, the open-source model ecosystem is exploding. Llama, Mistral, DeepSeek, Qwen — these models are getting better fast. But your subscription only gives you what their provider decided to carry. You don't get to choose.
The market is fragmented. Each provider has a slice of the models. Each has their own pricing scheme. Each has their own outages. None of them are solving this for you.
The Insight
Here's what's happening behind the scenes: the infrastructure running these models has excess capacity — idle compute that providers sell at a discount. Same models. Same hardware. Just underutilized.
At the same time, no single provider runs every open-source model. Each one picks a few. Each has their own outages. None of them are solving the fragmentation — because solving it would mean admitting you need more than them.
Two problems, one solution: aggregate the excess capacity from all of them, route requests intelligently to whoever has availability, and give the user every model under one roof.
The Product
routed.sh is a single subscription that connects you to multiple infrastructure providers simultaneously.
When you make a request, we find the provider with the best availability and lowest latency for your model. If that provider is slow or down, we reroute to the next one — instantly, invisibly. You don't write fallback code. You don't switch tabs to check a status page. You just keep working.
Three plans — Pro, Ultra, Max — with 100, 1,000, and 2,500 requests. Transparent limits. No hidden overages. Model costs are laid out in our docs so you always know where your requests go.
The API is OpenAI-compatible. Change your base URL and you're done. Same code. Same SDK. Zero migration.
We route your request to the right provider, at the right time, at the right price.
The Value
Better uptime than any single provider can offer.
A single provider is a single point of failure. We route across multiple. You only experience downtime if all of them are down simultaneously. Other subscriptions leave you waiting. We route around the problem.
The best open-source models, in one place.
The open-source ecosystem is moving fast. Each provider runs a few models. We bring them together. As we add partners, the catalog grows, and you get access automatically.
Lower cost because of how we buy.
We purchase excess capacity at a discount and pass the savings on. Same models, same outputs, a lower monthly price than you'd pay going direct.
A subscription that actually makes sense.
No token math. No surprise overages. You pick a plan, you know your limits, you use your requests across any model we offer. Model costs are transparent in our docs.
One API key, zero lock-in.
One key, one endpoint, every model we carry. And because we're OpenAI-compatible, you're never locked in — swap your base URL, keep your code.
The Positioning
We're not another coding AI subscription with one set of models and one set of servers. We're the layer on top of them.
We don't compete with providers on models — we carry them. We don't build GPUs — we use the ones already sitting idle. Our advantage isn't owning infrastructure. It's routing intelligently across infrastructure that already exists.
We focus on open-source models because that's where the momentum is. The gap between open and closed is shrinking every month. The ecosystem is rich, it's growing fast, and it deserves a subscription built for it — not a subscription that locks you into one provider's choices.
The Invitation
The name says it: routed. That's the whole product. We route your request to the right provider, at the right time, at the right price. You get better uptime, a growing model catalog, and a lower monthly cost — not because we built something new, but because we connected what's already there.
One subscription. Multiple providers. Better uptime. Start routing.
Every top model. One endpoint. Lower cost. Stop managing multiple provider accounts. Stop overpaying for inference. Start routing.
Get Started