Start with human intent
The product direction replaces configuration screens with conversation: describe the outcome, constraints, and duration instead of programming devices by hand.
OPEN SOURCE · AI-NATIVE · DEVELOPER PREVIEW
AetherIoT is building this experience: an agent turns your request into a home automation you can inspect and approve, then AetherEdge runs it locally under device capabilities, permissions, and safety rules.
AVAILABLE NOW: LOCAL RUNTIME, RULES, ALARMS, AND SAFETY INTERLOCKS · IN DEVELOPMENT: CONVERSATIONAL HOME SETUP
Illustrative data—not a statement of current device compatibility. Conversational home setup is still in development.
IN THE TARGET EXPERIENCE, YOU COULD SAY“After 10 p.m., turn off the living-room light when nobody has been there for 10 minutes, set climate to 26°C, and only notify me if the door is unlocked.”
After 22:00 + empty for 10 min → lights off / climate 26°C / notify if unlocked
WHY AETHER
Traditional automation asks people to translate life into entities, triggers, conditions, and actions. Aether is building the governed runtime foundation for agents to do that work safely.
The product direction replaces configuration screens with conversation: describe the outcome, constraints, and duration instead of programming devices by hand.
Typed capabilities, revisions, permissions, confirmation, and audit turn model output into inspectable artifacts—not direct device calls.
AetherEdge runs commissioned acquisition, rules, alarms, history, and safety locally, even when the agent, cloud, or internet disappears.
THE OPERATING PRINCIPLE
“Agents propose home automations. Permissions and safety rules check them. The in-home edge runtime executes them.”
ONE AI-NATIVE SYSTEM
The beta already provides deterministic edge execution, governed application boundaries, public contracts, agent-readable docs, and cloud foundations. Conversational generation, simulation, and continuous adaptation remain explicit product direction.
DETERMINISTIC EDGE
The Linux runtime that owns live state and executes commissioned behavior without placing an LLM in the real-time loop.
Explore the edge runtime 02AGENT AND CONTROL PLANE
The evolving home for agent context, desired state, governed jobs, integrations, and provider-native cloud coordination.
See the cloud foundation 03TYPED INTEROPERABILITY
The language-neutral source of truth that keeps capabilities and messages type-safe, versioned, portable, and testable.
Inspect the contractsThe first domain solution built on AetherIoT, supplying energy semantics and safety constraints without changing platform authority.
ONE DOCUMENTATION HOME
Follow the AI-native architecture, inspect current capabilities, and verify exactly which pieces are implemented, experimental, or planned.
See how conversation becomes a governed plan and deterministic edge behavior.
OverviewProduct boundaries, authority, deployments, and user journeys.
AetherEdgeInstall and build with the edge runtime, Kernel, CLI, and SDK.
AetherCloudUnderstand the cloud fusion and governed control plane.
AetherContractsUse the shared protocol, schemas, fixtures, and TCK.
Edge–Contracts–Cloud integration guideFollow the complete AetherEdge → AetherContracts → AetherCloud integration flow.
CompatibilityChoose tested product and contract version combinations.
RoadmapSeparate implemented, experimental, and planned capabilities.
FOUNDATION AVAILABLE TODAY
Start with a safe-empty edge runtime. Let an agent inspect typed capabilities and documentation before it proposes any change.
Read the documentation$ git clone github.com/EvanL1/AetherEdge
$ cd AetherEdge
$ cargo run -p aether-example-minimal-gateway
# local composition · no broker · no cloud · no deviceTHE INTERFACE BECOMES A CONVERSATION
Build the open foundation that lets agents generate change while contracts, policy, and deterministic edge execution keep it real.