Roadmap
Roadmap
Three stages, one direction: from secure inference to your models doing the work. Last updated July 1, 2026
Secure & private inference
Now- Region-pinned inference on dedicated capacity — EU, USA, LATAM
- No storage, no sharing, no training on your data
- Qwen 3.8 (320K context) under the 99.999% SLA
- Starter, Business, Advanced and Enterprise plans
Knowledge, memory & connection
In development- Brain — Enterprise Knowledge: 10 / 20 / 50 GB by plan, Unlimited on Enterprise
- Knowledge bases
- Memory across sessions
- Connectors
- Knowledge Intelligence
- Advanced Guardrails
- MCP & tool access
Your models, your agents
Planned- Bring your own models (BYOM)
- Dedicated, customer-specific models
- Dedicated infrastructure
- Agent orchestration
How to read this roadmap
Stages follow the approved product direction. V1 is live in each region as capacity lands; In development and Planned are not live today and carry no date commitments. We re-verify every stage against what actually ships.