# How to use Eigent, and where an open source team runs it

> Eigent is an open-source Cowork desktop app: clone the repo, install the dependencies, run it locally, and point it at a model you choose. It runs one workforce on one machine, against your own files. Kortix, the open-source AI Operating System, is where a team takes the same work: one repo, agents on cloud computers, and reviewed changes.

URL: https://eigentalternative.com/how-to-use-eigent

## Eigent's own commands, then Kortix's three

The install and dependency steps from Eigent's repository, beside the Kortix setup.

### Eigent preview

`terminal`

```sh
$ git clone https://github.com/eigent-ai/eigent.git
$ cd eigent
$ npm install
$ npm run dev
```

### Eigent deps

`terminal`

```sh
$ npm install
$ cd backend
$ uv sync
```

### Kortix

`terminal`

```sh
$ curl -fsSL https://kortix.com/install | bash
$ kortix init
$ kortix ship
```

## What Eigent actually is

Eigent is an open-source Cowork desktop application. Its repository describes it as a local-first app that builds, manages, and deploys a custom AI workforce, built on CAMEL-AI's multi-agent framework and packaged as an Electron desktop app with a FastAPI backend ([Eigent on GitHub](https://github.com/eigent-ai/eigent)). You run Eigent on your own machine. The quick-preview commands in its repository connect to Eigent's cloud services and ask you to register an account first.

## The real setup path

Two ways in. The quick preview clones the repository and runs it with npm, and it depends on Eigent's cloud backend. For a standalone setup the repository points to Local Deployment, which includes a local backend server with the full API and local model integration through vLLM, Ollama, or LM Studio, with no cloud account required ([Eigent on GitHub](https://github.com/eigent-ai/eigent)). Either way, you choose the model yourself and keep the credentials on the machine.

## Picking a provider

Eigent is model-agnostic. You connect a cloud API provider, an enterprise gateway, or a local model, and bring your own key, then switch providers without rebuilding a workflow. Its product page lists 16 supported models and offers free self-hosting with your own keys or local models ([Eigent](https://www.eigent.ai)). This step decides what your agents can reach, because a local model keeps the work on the machine while a cloud provider sends it out.

## Spaces, agents, and a first task

Work in Eigent is organized into Spaces. A Space packages the agents, Skills, Connectors, and instructions a team reuses, so a new Space starts with the setup already in place, and memory carries from one task to the next, scoped to you, the Space, or the session. A first task is a folder plus a goal. The repository's own examples include reviewing a month of pull requests with a locally hosted model and drafting a Slack update, or organizing a cluttered desktop. You watch each agent's progress, then approve the result or ask for revisions.

## Where one desktop stops serving a team

A desktop install is one workforce on one machine, per person. That is the limit of the setup: the configuration lives on that computer, so a team has no single shared config to review, no one place where every change lands, and no session that keeps running after the laptop closes. Eigent's managed cloud tier exists for teams that want hosted infrastructure, but the app model is still a desktop workforce.

## Where a team runs the same work

Kortix holds the same work for a whole team. Your agents, skills, company memory, and connectors live in one git repo the team owns, and every session runs on its own cloud computer, on its own branch, so thousands can run in parallel on one config. Work reaches `main` through a change request a person reviews, and merge is default-deny for agents. You run it on Kortix Cloud, in your VPC, or on your own on-prem network, and self-host is free ([Kortix docs](https://kortix.com/docs)).

The [self-hosting guide](/self-hosting-eigent-alternative) covers running the platform on your own box, the [Kortix vs Eigent](/kortix-vs-eigent) page compares the two platforms directly, and [Eigent pricing](/eigent-pricing) covers what Eigent charges. To run the whole system yourself, [Try Kortix](https://kortix.com).