Skip to content

The bigger picture. ​

A private AI workspace, built to put you in control. Each milestone contains smaller releases. The order describes the broader direction, not a dated delivery schedule.

Together, these milestones form a private AI ecosystem: your workspace, your personal intelligence, agents working with your permission, and a sovereign execution layer for people and developers. Read the full vision.

01. Private AI foundation ​

Available · expanding. A dependable place for everyday work, with meaningful control over data, models and cost.

  • Available: Local and cloud AI, local history and optional memory.
  • Available: ZKR for text and supported reviewed files.
  • Available: Cloud model choices, wallet credits and usage.
  • Available: Refined workspace and unified model discovery.
  • Next: Easier onboarding.
  • Available: Temporary chats, reusable prompts and local memory controls.

02. Multimodal creation ​

Available · expanding. Turn an idea into something you can use, refine and keep.

  • Available: Linked code previews, editing, saved versions and project ZIP downloads.
  • Available: Image generation and reference editing.
  • Available: Silent video generation, speech and audio transcription.
  • Available: Local media Library and downloads.
  • Planned: Richer coding and execution options.

03. Connected knowledge ​

Planned. Keep the context, sources and outputs of your work together.

  • Available: Projects that organise conversations.
  • Available: Shared project instructions and selected file context.
  • Available: Optional web search with source links.
  • Planned: Longer documents and better context management.
  • Planned: Sourced research and connections to your information.
  • Planned: Local backup, import and recovery.

04. Agents and automation ​

Available · expanding. Delegate useful work with clear permissions, spending limits and a way to stop.

05. Open developer ecosystem ​

Available · expanding. Bring Noema into the tools you use and build new experiences around it.

06. Sovereign GPU compute ​

Final milestone. A sovereign execution layer connecting personal hardware and verifiable confidential compute, with privacy policies that travel with the task.

  • Planned: Privacy-aware routing across local and confidential execution.
  • Planned: First confidential GPU service and hardware attestation.
  • Planned: Independent relay and inference settlement.
  • Planned: Wider operators and onion routing.
  • Planned: Operator tools and defined network incentives.

ZKR addresses detected content. Confidential hardware addresses execution. Onion routing addresses network metadata. The current cloud Router does not provide the planned Network, and the live token does not make those services available.

A private AI workspace, built to put you in control.