We have a new partner. Requesty, the AI gateway for teams building and scaling AI products, is connecting Melious to its platform. Teams that route their traffic through Requesty can now run open-weight models on Melious, on hardware in European datacenters.
Who Requesty is
Requesty connects applications to more than 600 model endpoints through one API, with routing, cost management, observability and governance built in. The platform handles 225B+ tokens every day.
When a provider goes down, traffic switches to the next best option in under 14 milliseconds, and Requesty reports 99.99% uptime. EU traffic stays in Frankfurt, personal data is scrubbed before a request reaches a provider, and every call shows up in cost and latency analytics.
What we bring to the gateway
Requesty decides which model handles which request, at what cost, under which rules. What a gateway cannot decide on its own is where the compute physically runs. That is where Melious comes in.
We combine datacenters across Europe into one platform and run the best open-weight models in the world on it, faster and cheaper than US hyperscalers, without compromises on privacy. On our servers a model is just a set of weights. It cannot send data back to its developer, whether it was trained in Europe, the US or China. Customer data is never stored and never used for training.
Because it is Melious underneath, every call runs on 100% European infrastructure, so your data never leaves Europe. With up to 250 tokens per second, you get the best price/performance in the EU, running on low-carbon energy.
Why this matters
AI is becoming as indispensable as electricity. Almost all of it runs through a handful of companies outside Europe, and anyone who wanted a European setup had to settle for the second-best option. That is the gap this partnership closes. Requesty users keep the routing, cost control and oversight they already have. Behind it, their inference runs in Europe, under European law, outside the reach of the US CLOUD Act.
Rollout
Melious is live in Requesty starting today. As of today, you can route GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash and DeepSeek V4 Flash through the same endpoint and SDK your team already builds on, from $0.12 per 1M tokens. All of it runs on European datacenters, from partners like OVHcloud, Scaleway and Verda.
The same platform also runs our coding subscription, so coding sessions get the same treatment. European hardware, Zero Data Retention, no data leaving the EU.
Run open models on European ground
Open-weight models. Zero Data Retention. Inference in European datacenters.