Chat with open models, and choose where they run
Every model states where your messages go. Sign in with your Piatra Institute account.
Sign inYou sign in on account.piatra.institute and return here.
Three places a model can run
- 01
Piatra Center
Our cluster, always on.
Your messages stay in the Piatra Center cluster.
- 02
Private GPU
A GPU started only for you on RunPod in the EU, billed per second while it runs. It stops after 10 idle minutes.
Your messages go to a GPU started only for you, on RunPod in the EU.
- 03
Hosted
Third-party providers, reached through OpenRouter.
Your messages go to a third-party provider through OpenRouter.