# Claude Cowork for teams, and the open-source platform a team owns

> Claude Cowork is included on Claude's Pro, Max, Team and Enterprise plans, and it runs inside Anthropic's cloud. This guide covers what a team can and cannot own with it, how a team governs agent work, and where open-source Kortix fits.

URL: https://opensourceclaudecowork.com/claude-cowork-for-teams

_Quickstart_

## The company, as files in one repo

`acme-co`

```
acme-co/          # your whole company, as files
  agents/         # who does the work
  skills/         # how your company does a job
  memory/         # what it has learned so far
  kortix.yaml     # rules, triggers, connectors, machine image
```

Source: kortix.com/docs

A Kortix project keeps the agents, skills, memory and connector config in a git repo the team owns.

## What a team owns when it moves past Claude Cowork

Claude Cowork is included on Claude's Pro, Max, Team and Enterprise plans, and it is built for non-coding knowledge work inside Claude's app. A team can use it today, and Anthropic gives Teams and Enterprise admins real controls: access set by team, spend limits, tool permissions by department, and activity streamed to your SIEM through OpenTelemetry. What the team does not get is ownership. Anthropic publishes the product, not the source behind it, there is no self-hosted edition, and the configuration and work history stay in the Claude account. This guide covers what a team keeps when it decides that is not enough.

### The repo is the thing a team owns

Kortix is the open-source AI Operating System: your agents, their skills, your company memory, and every connector in one git repo you own. It is the leading open-source alternative to Claude Cowork and ChatGPT Work. An agent is `agents/<name>.md` (behavior) plus an `agents.<name>` block in `kortix.yaml` (governance). The `.md` is a stock OpenCode agent file, and the scoping field is `permission`. Because every part is text, a team can `grep` the rules an agent runs under, diff a change to a skill, and roll any part of it back. That is the ownership a hosted product cannot offer, whatever admin controls it ships.

### Governance a team can read

A hosted agent asks a team to trust a control panel. Kortix puts the controls in the repo and in live project state. Per-resource permissions for people and agents. Roles, groups, and an audit trail. SSO is SAML 2.0 single sign-on and SCIM 2.0. Approval gates you set. Off until you set them. Secrets are encrypted at rest with a key per project, and connector credentials are brokered server-side and never enter the machine. Each session has its own isolated machine and branch. Session work reaches `main` through a change request. Merge is default-deny for agents, so a person reads the diff before anything lands.

### Self-hosting a team's platform

Self-hosting is where ownership stops being abstract. Run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free, and one Docker Compose stack carries the platform on a Linux box the team controls. The [self-hosting guide](/self-hosting) covers the two commands, the domain and ports, and the three directories to back up. Any model provider with your own keys, so the model stays a choice the team makes. The [open-source Claude Cowork alternatives](/claude-cowork-alternatives) page maps the wider field, and [Read the Kortix blog](https://kortix.com/blog/kortix-vs-claude-cowork) for the same-company comparison.

_FAQ_

## Claude Cowork for teams, the questions teams ask

### Can a team actually use Claude Cowork?

Yes. Claude Cowork is included on Claude's Pro, Max, Team and Enterprise plans, and on Teams and Enterprise an admin can set access by team, set tool permissions by department, and stream activity to a SIEM. What a team cannot do is own the source, run it on its own hardware, or keep the configuration outside the Claude account.

### What does a team own when it runs Kortix?

One git repo. Agents and skills are markdown, memory is files, and `kortix.yaml` declares the machine image, the connectors and the triggers. The repo is versioned, diffable and owned outright, so a team can read every rule its agents run under and roll any part of it back.

### How does a team govern what agents do?

Per-resource permissions for people and agents. Roles, groups, and an audit trail. Approval gates you set. Off until you set them. Session work reaches `main` through a change request. Merge is default-deny for agents, so a person reads the diff and decides what lands.

### Can a team self-host Kortix?

Yes. Run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free, and the [self-hosting guide](/self-hosting) covers the two commands. A team needs one persistent domain, a sandbox provider key, model keys in the dashboard, and named reviewers for change requests.

## Run your team's agents on a repo you own

Start with one job, then invite the team.

[Try Kortix](https://kortix.com) · [Self-hosting](https://opensourceclaudecowork.com/self-hosting)

Open source · Any model, your keys · Self-host, VPC, or on-prem