Skip to main content
elan1 Run

Agentic AI runtime

It runs where you say.

elan1's own runtime, on one signed image, in the cloud, your data centre or an appliance. It refuses before it acts. It proves what happened.

Book a demo
run.elan1.ai/fleet

1 signed image · fleet

  • Edge · Salem plant

    Healthy

  • Department · Finance, Chennai

    Healthy

  • Rack · DC-1

    Healthy

  • Cloud · ap-south-1 Mumbai

    3 zones
Image signature verified
Fleet screen for one signed image: an Edge appliance at the Salem plant, a Department appliance for Finance in Chennai and a Rack in DC-1, all healthy, plus the cloud in ap-south-1 Mumbai across 3 zones. The image signature is verified. Illustrative data.

What is elan1 Run?

elan1 Run is elan1's own agentic AI runtime, on one signed image, in the cloud, your data centre or an Edge, Department or Rack appliance. Every elan1 product and every agent action runs on its nine contracts. It routes and serves models, runs agent tools in a sandbox and traces every call, while elan1 Govern decides which actions happen.

Updated

The challenge

Where AI runs is asked first.

Boards, regulators and customers now ask where models, data and agents live before they ask what they do. Most stacks cannot answer in one sentence.

77% of organisations weigh a vendor's country of origin. Deloitte, 2026

A site engineer traces a fibre cable at an open rack in a dark equipment room

  • 01

    Parts from everywhere.

    Every component is a separate answer.

  • 02

    Data that wanders.

    Prompts and records cross borders unseen.

  • 03

    No proof of the run.

    Nobody can show what actually executed.

How it works.

  1. 01

    Choose where.

    Cloud, your data centre, or an Edge, Department or Rack appliance.

  2. 02

    Install one signed image.

  3. 03

    Agents run on nine contracts.

  4. 04elan1 Govern

    Govern decides each action.

    Run refuses before it acts.

  5. 05

    Every run is traced and proven.

Take a look inside.

Nine contracts underneath every agent.

  • run.route
  • run.serve
  • run.flow
  • run.tools
  • run.sandbox
  • run.index
  • run.trace
  • run.identity
  • run.fleet
run.elan1.ai/contracts

Contracts

9 · one signed image

run.route     run.serve     run.flow
run.tools     run.sandbox   run.index
run.trace     run.identity  run.fleet

Every elan1 product and every agent runs on the same nine

Contracts screen listing the nine contracts on one signed image: run.route, run.serve, run.flow, run.tools, run.sandbox, run.index, run.trace, run.identity and run.fleet. Illustrative data.
Take the tour

By the numbers.

  • 9

    contracts

  • 3

    appliance sizes

  • 1

    signed image

  • 3

    availability zones in Mumbai

    CO-01

An engineer slides a server unit out of a rack with both hands in a data hall with slate grey cabinets

Where it works.

All industries

How it fits

Run carries. Govern decides.

Run carries, routes and records every call. It never decides which actions happen. That is elan1 Govern's job.

Applications

elan1 Actelan1 Propelelan1 Resolveelan1 Pulse

Foundation

elan1 Ontologyelan1 Decideelan1 Governelan1 Run
elan1 Run on the elan1 platform.

Questions

elan1 Run, answered plainly.

Ask us anything

An agentic AI runtime is the layer agents run on. It routes and serves models, runs agent flows and tools in a sandbox, and traces, identifies and manages the fleet. elan1 Run does this on one signed image with nine contracts, from run.route to run.fleet. It carries every action but never decides which actions happen: elan1 Govern does.

AWS Mumbai by default, with no US region. It can also run in your data centre or on an appliance.

Models from Anthropic, OpenAI, Google and open source, behind run.serve.

Below about 1 billion tokens a month, usually yes, and we will tell you so.

There are three: Edge, Department and Rack. Each runs the same signed image as the cloud, and every appliance is managed as part of one fleet, with every call traced and every agent given an identity. Appliances are for sites that must keep their data inside their own walls, such as a plant, a finance department or a data centre.

You can build an agent. The hard part is around it: rules that refuse by default, approvals that expire, credentials the agent never holds, and receipts your auditor verifies without trusting you. That part is the product.

Start with one process.

Name the place. We'll show you the run.

Book a demo
  • Allowed · Approved by Asha R.
  • Held for CFO
  • Refused · rule 4.2
Illustrative data.

Start with one process.

Add the next when it holds.

Book a demo
Approvals · Kolam ComponentsIllustrative
  • Allowed · Approved by Asha R.
  • Held for CFO
  • Refused · rule 4.2
RCPT-0042 · signed · chained
A product view with three decisions: allowed and approved by Asha R., held for the CFO, and refused under rule 4.2. Illustrative data.