Rutli status
Ambitious enough to say where we're going. Serious enough to say where we are.
No fake scale. No pretend public market. No imaginary capacity. Everything gets a label, and every label on this site comes from the same list as this page; if a page and this page disagree, the build fails. Last built 2026-09-06.
The founding proof, capability by capability.
Can independently owned, heterogeneous machines cooperate without ceasing to be independent? Three machines, and a list that turns to LIVE one item at a time.
- 0001 A GPU server in the founder's house running LIVE
- 0002 A 128 GB Mac on the same network running LIVE
- 0003 A collaborator's Mac, behind a home router joins when the node app ships PLANNED
- 0004 A node in Italy open PLANNED
None of these machines belongs to Rutli. Rutli coordinates only the capacity each owner chooses to share.
- Node registrationPLANNED
- Capability and availability reportingPLANNED
- Remote workload execution across nodesToday: inference routed between two machines with fallback. The general case is planned.PLANNED
- Metering and correct owner attributionPLANNED
- Contribution on/off, schedules, resource limitsShown as it will look; not connected yet.PLANNED
- Local priority on the founding machinesA task of the owner's preempts the other machine's request; proven on nodes 0001 and 0002.LIVE
- Private plane and network plane isolationPLANNED
- The contribution dashboard and the balancePLANNED
LIVE. Running now, demonstrable.
- ALYTHome intelligence on a real house: nine connector kinds, a Zigbee network, routines, a voice hub, an Alexa add-on, a phone app.LIVE
- FinanlyRuns real companies' books daily: BI, close, reconciliation, self-hosted with local AI.LIVE
- OrganikoUSDA organic compliance and traceability workflows.LIVE
- Kayro Messenger and Kayro CloudA messenger with calls and a cloud client, rebuilt, on the founder's own Matrix and Nextcloud.LIVE
- Local AI appsChat, search with citations, deep research, image generation, voice in and out, all on hardware in one house.LIVE
- Two machines routing inferenceA GPU server and a 128 GB Mac route inference between them with fallback and local priority, proven 2026-08-24.LIVE
- Local voice, physical AIA room-level voice hub answering from intelligence in the house; runs in the founder's home daily.LIVE
- Private collaboration stackSelf-hosted communications, files and local AI on Matrix and Nextcloud, with Kayro on the phone.LIVE
PILOT. Working in controlled early use.
- Founding NodesOpens when the node application exists for Windows, Linux and macOS.PILOT
PLANNED. Committed direction, not yet operating publicly.
- The node applicationWindows, Linux and macOS from the start.PLANNED
- Registration, discovery, remote execution, meteringThe network plane.PLANNED
- The owner's rulesLocal priority, hours, ceiling, power cap, workload classes, jurisdiction.PLANNED
- One dollar in, one dollar to the ownerThe principle is stated; the network that pays it does not run yet.PLANNED
- Published prices per model, and the owner's price per unit of workReference prices for the alpha; set with the founding nodes; every price public; the compute price goes 100% to the node owner.PLANNED
- Frontier open models served across many owned machinesKimi K2.6, DeepSeek V3.2, Qwen3 235B, GLM-5, MiniMax M2.7: models no single desk can hold, run across the network. Needs the network plane.PLANNED
- External Founding Node cohortOpens when the node application exists for Windows, Linux and macOS.PLANNED
- Public compute market and buyer interfacePLANNED
- Public developer APIPLANNED
RESEARCH. Under investigation, not a product commitment.
- External workload marketplaceRESEARCH
- Distributed trainingRESEARCH
- Regional sovereign poolsRESEARCH
- Distributed inference across many nodesHow the big open models would be served across owned machines.RESEARCH
- Federated learning, storage, other network servicesRESEARCH
Why publish all this?
Because infrastructure should not require faith. Participants should know what exists, what does not, what the node is doing, what the network can access, how the economics work, which rules are stable, and which are still changing.
Transparency is part of the product.