Models
You choose which model runs your workload. Qwen 3.8 with a 320K context window runs on dedicated servers; the lineup grows through the roadmap.
Product
Platform
One control layer over your models, your data, your knowledge and your region.
The control layer, in three pillars
Data & knowledge
Your prompts, your data and your knowledge are not stored, not shared and not used for training. Memory — the knowledge pillar — arrives with V2: private to your workspace, pinned to your region.
Region
EU, USA or LATAM: your workloads run in the region you choose, and your endpoints name it.
Explore the platform
Start with control: join the waitlist and pick the region your data must stay in.