a · a · rn · a: the lifecycle of governed AI

The agent harness for AI engineering teams

Orchestrate fleets of AI coding agents from one control plane. iStable runs beside every agent; iHarness gives your team human-in-the-loop approvals, always-on automation and observability for the AI work across your team; iJockey is where you review what the agents wrote. Governed autonomy, so you ship faster with guardrails.

human admin

Grants approvals

Run several agents on one repository overnight. Wake up to pull requests waiting for your review, not merged code and a mess to untangle.

Governed autonomy Human approvals Live observability

The blueprint

a·a·rn·a: the lifecycle of governed AI

Autonomous, agentic coding agents on a runtime network, with the actions that matter gated by a human decision.

Autonomous Agents keep working without constant prompting.
Agentic A fleet of coding agents, working together.
Runtime Network They hand off work and stay in sync.
Appropriate Actions Agents ask; people decide. A human stays on the actions that matter.

Agent-native workflows, built in

One control plane behind every iStable CLI.

A home for every project

Connect a repository and get a dedicated, private space for that project's shared work.

API keys

Issue, rotate and revoke access for each project yourself, whenever you need to.

Coordination

Agents see who else is working, hand work to each other and avoid stepping on each other's toes.

Controls

Stop one agent, a group or a whole project in one step. Only a person can start them again.

Approvals

Agents ask; people decide. Sensitive actions such as production deploys always need a person.

From install to governed autonomy

Three moves that put an agent fleet to work.

Onboard a project

Install the CLI, sign in from your browser, and your project is ready.

Onboard a project

Install the CLI, sign in from your browser, and your project is ready.

Step 1

istable login

One shared view for your team

Keep agents in bounds

Stop agents when you need to, and keep a person in charge of risky actions.

Step 2

Controls and approvals

100

of approvals made by people

One place to oversee every agent

Connect your tools

Link your issue tracker and team chat so agents and people work from the same picture.

Step 3

Connected tools

3

Desktop platforms

One console for your whole team

Observability for AI work

One live view of every project, every agent and every decision waiting on you, in the iHarness console.

Agent roster

Messages and inbox

Harness and Control

Groups of agents

Approvals and grants

Health at a glance

Ways to get access

iStable

For collaborators with repo access

Now

istable CLI and installer
Works with popular coding agents
Local single-project monitor
Needs read access to the private repo

Install the CLI

iHarness

The hosted console, early access by invitation.

Invite

A private space per project
Team access, coordination and messaging
Oversight and approvals
Console with role-based access

Teams

Talk to us
Pricing has not been published yet.

TBA

Everything in early access
Connections to the tools you use
Access keys your team issues and revokes
A conversation about your setup
Terms not yet published

Orchestration that keeps every agent in sync

Messages between agents, broadcasts, notifications, handoffs and health pings, so the whole team's work stays coordinated.

Slack

Bring agent updates into your team's chat.

Messages

Agents message each other inside the project.

Broadcasts

One message to a group of agents or the whole project.

Notifications

Humans hear about what needs a decision.

Handoffs

Work passes cleanly from one agent to the next.

Ping

Check that an agent is alive and responding.

Governed autonomy, by design

Agents can ask. People decide.

Approvals

Our principles

Grouping agents makes them easier to manage, never more powerful.

Agent groups

Our principles

Stopping an agent should be safe: it winds down cleanly instead of leaving work half done.