A bigger picture.
Six milestones. Smaller, meaningful releases along the way. From local intelligence to a private, independent AI ecosystem. More capability, with your control carried through every phase.
Every milestone adds useful capabilities while strengthening ownership, privacy and independence. This is the broader direction, with smaller releases inside each phase.
Private AI foundation
Anonymous entry, local intelligence and conversations that stay with you. Build the private foundation for everything that follows.
- AvailableLocal and cloud AI, local history and optional memory
- AvailableZKR for text and supported reviewed files
- AvailableCloud model choices, wallet credits and usage
- AvailableRefined workspace and unified model discovery
- NextEasier onboarding
- AvailableTemporary chats, reusable prompts and local memory controls
Multimodal creation
Turn an idea into code, images, video or sound. Keep your work in a local library, with the freedom to download it and take it with you.
- AvailableLinked code previews, saved edits and ZIP downloads
- AvailableImage generation and reference editing
- AvailableSilent video, speech and audio transcription
- AvailableLocal media Library and downloads
- PlannedRicher coding and execution options
Connected knowledge
Bring your information together without giving up control of it. Research, sources and local projects that stay yours.
- AvailableProjects for organising conversations
- AvailableShared project instructions and selected file context
- AvailableOptional web search with source links
- PlannedLonger documents and better context management
- PlannedSourced research and connections to your information
- PlannedLocal backup, import and recovery
Agents and automation
Delegate useful work with clear permissions, spending limits and a way to stop.
- PlannedReusable assistants and multi-step tasks
- AvailableBrowser task permissions, approvals and visible progress
- AvailableAn optional desktop extension for work across permitted sites
- LaterDesktop options only where they add useful capabilities
Open developer ecosystem
Give applications and agents the same control over models, privacy and spending as the Noema workspace.
- AvailableText developer API with revocable keys, budgets and model permissions
- PlannedPublic Engine delivery and integrations
- AvailableText streaming and per-key usage
- PlannedSDKs and tool calling
- PlannedPortable execution receipts
Sovereign GPU compute
A sovereign execution layer connecting personal hardware and verifiable confidential compute, with privacy policies that travel with the task.
- PlannedPrivacy-aware routing across local and confidential execution
- PlannedFirst confidential GPU service and hardware attestation
- PlannedIndependent relay and inference settlement
- PlannedWider operators and onion routing
- PlannedOperator tools and defined network incentives