Skip to main content

runkora docs

The open source control plane for running agent factories in production.

Deploy runkorarunkora

01 · Agent execution

How does an agent run?

02 · Identity

Who is the agent?

03 · Policy

What stops it from doing damage?

04 · Connectivity

What can it talk to?

Getting started

01 · For the platform team

Deploy runkora

Get kora running on your cluster ready for onboarding your first team.

Install kora
  1. 1

    Install kora

    Install the CRDs chart, then kora with your values file.

    Read →
  2. 2

    Upgrade kora

    Move to a new version and check what changed.

    Read →
  3. 3

    Observability

    Find the logs, metrics and traces kora emits.

    Read →

02 · For platform and team leads

Set up a team

Set up your teams and permission sets to get started building agents.

Onboard a tenant
  1. 1

    Onboard a tenant

    Give a team its own namespace and identity.

    Read →
  2. 2

    Connect model providers

    Choose which models the team’s agents can use.

    Read →
  3. 3

    Grant access

    Decide who can create agents and start runs.

    Read →

03 · For agent developers

Build an agent

Get an agent running on the platform.

Create an agent
  1. 1

    Create an agent

    Define an agent type and what it’s allowed to do.

    Read →
  2. 2

    Environments

    Set up the sandbox its runs work in.

    Read →
  3. 3

    Trigger runs

    Start runs from a pull request, an alert or a schedule.

    Read →

04 · For whoever owns the system

Connect a tool or system

Give your teams and agents the ability to connect to your systems.

Connections
  1. 1

    Connections

    Link an external system like GitHub or Slack.

    Read →
  2. 2

    MCP servers

    Give agents tools through the tool gateway.

    Read →
  3. 3

    Egress

    See how every outbound call is checked against policy and given a short-lived credential.

    Read →

Decisions

We track all key decisions for our platform in the form of PRDs and ADRs. To see how we got to where we are, you can read through our decision log.

All decisionsHow do we write them?