PLATFORM

One operating layer between AI and your operations.

Relayn OS watches what happens in your systems, keeps a working picture of your organisation, and lets AI act only within rules your people have approved. We call this execution control.

01 / How it works

Observe. Understand. Act.

Relayn OS observes, understands and acts, and outcomes feed back.
StageWhat happensRead more
ObserveRelayn takes in what your operation already produces: sensor readings, camera images, documents, email, and the screens of older systems.What it connects to
UnderstandIt stores what it learns as facts with dates. It knows which pump, which limit, who is on call tonight, and what changed last week.How memory works
ActA language model works out what is being asked and proposes an action. Your rules decide whether it runs, waits for a person, or is stopped. The outcome is recorded.How control works

02 / What people do

For your people, it is a workspace.

Think of the AI workspaces people already know, built for one company and connected to that company's systems.

Ask

Get answers from your company's own knowledge and history, with the source.

Hand off

Give an agent a job: answer email, prepare a report, watch plant status.

Approve

Decide on the actions that need a person, from wherever you are.

Build

Describe a new agent in plain language, or bring one you already have.

See how people work with it

03 / What is inside

Seven parts, each with one job.

  • ControlDecides what reaches the real world: rules, approvals and a record of every decision.
  • ConnectionLinks to your systems, channels, cameras and sensors.
  • IntentTurns everyday language into requests that can be checked.
  • Knowledge and memoryHolds what your company knows and what is true now.
  • Ready workflowsAgent chat, email handling, storage and search from the first day.
  • BuilderThe tools to make your own agents, in chat or in code.
  • Open edgesYour models, keys, agents and AI tools plug in.

04 / What runs today

What runs today, and what is next.

Most of Relayn OS already runs for clients, set up by our team. We say plainly which parts are earlier.

Examples of how we label parts by maturity. Status applies per deployment; this is not a promise that every row is live for you on day one.

StatusParts
Running for clientsRules and control · knowledge and memory · intent · agents and ready workflows · older systems through their own screens · agents on phone calls · vision
Available on requestDeployment on your premises · your own models and keys
In pilotIndustrial sensors
Early accessBuilding agents in code · connecting your own agent or AI tool
Coming nextBuilding agents in chat for everyone