- Open source
- Yes. Read it, fork it, audit it.
- Where agents run
- Each session on its own cloud computer.
- Where the config lives
- In one git repo you own.
- Who can use it
- A team, from web, Slack, Teams or CLI.
- How work lands
- Through a change request a person reviews.
- Parallel agents
- Thousands of agents in parallel on one config.
Eigent on GitHub: how to run the open source repo
Eigent's code lives at github.com/eigent-ai/eigent, an open source Cowork desktop app you clone and run on your own machine. When one desktop repo stops carrying a team's work, Kortix is the open-source AI Operating System where agents, skills, memory and connectors live in one git repo you own.
Kortix and a local desktop repo, side by side
What changes when the same idea runs from one repo a whole team owns.
- Open source
- Yes. Clone, inspect and modify the code.
- Where agents run
- Your machine, local-first execution.
- Where the config lives
- On your machine; Spaces reuse it.
- Who can use it
- A desktop app; managed cloud and enterprise on offer.
- How work lands
- In your local files and Library.
- Parallel agents
- Parallel agents on one machine.
Clone Eigent and run it from source.
1$ git clone https://github.com/eigent-ai/eigent.git2$ cd eigent3$ npm install4$ npm run devThe repository README's commands; the preview connects to Eigent cloud, and local deployment runs standalone.
Where Eigent's repository lives
Eigent's code is at github.com/eigent-ai/eigent, and the repository is titled the Open Source Cowork Desktop, a local and free alternative to Claude Cowork and Codex (eigent-ai/eigent). The README describes Eigent as the open source Cowork desktop application, a multi-agent workforce built on CAMEL-AI that builds, manages and deploys a custom AI workforce. That is the repository developers look for when they search Eigent on GitHub.
How to run Eigent from source
Clone the repository, then run npm install and npm run dev with Node.js 18 to 22 installed. That preview path connects to Eigent cloud services and needs an account registration. For a standalone run, the README points to Local Deployment: a local backend server with the full API and local model integration through vLLM, Ollama or LM Studio, with no cloud account required. After a pull, backend dependencies update with cd backend and uv sync. The repository documents its stack as FastAPI, uv and Uvicorn on the server, CAMEL for multi-agent orchestration, and React, Electron and TypeScript for the desktop app.
What the desktop app runs
Eigent's README documents a single-agent harness for focused tasks and a multi-agent mode where agents divide work and run in parallel, a built-in browser and terminal toolkit, MCP and skill integration, and scheduled recurring workflows. It is model-agnostic, connecting cloud APIs, an enterprise gateway or local models. Agents run on your machine with local-first execution, so your files, credentials and context stay on it. The eigent.ai site adds Spaces, which package agents, skills and connectors once for reuse, and Memory Carry, which carries confirmed knowledge from one task to the next (eigent.ai).
When one desktop repo stops being enough
A desktop repository keeps its configuration on the machine that runs it, which suits one person at one desk. A team needs the agents, the memory and the connector setup to be shared, versioned and reviewable in one place. Kortix answers that with a project that is a git repo: kortix.yaml sits at its root, and agents, skills, memory, tools, triggers and connectors are files in it (Read the docs). Kortix hosts the repo, or you connect your own, and each session boots its own isolated sandbox on a branch named after the session.
From a clone to a repo a team owns
kortix init scaffolds a project with kortix.yaml and your agents and skills, and kortix ship pushes the repo and brings it live (Kortix on GitHub). The agent's work reaches main only through a change request a person approves, so nothing lands unreviewed, and self-host is free on Kortix Cloud, in your VPC, or on your own on-prem network. A team that still wants a local-first desktop app has other open-source options on this site (Eigent alternatives), and the wider open-source field is compared by Kortix on self-hosting, model choice and review.
Questions developers ask about Eigent's repo
Short answers on where it lives, how it runs and what comes next.
Where is Eigent's GitHub repository?
Eigent's code is at github.com/eigent-ai/eigent, titled the Open Source Cowork Desktop. The repository holds the multi-agent application and its documentation.
How do I run Eigent locally from source?
Clone the repo, install Node.js 18 to 22, then run npm install and npm run dev. For a standalone run, the README's Local Deployment path uses a local backend and a local model through vLLM, Ollama or LM Studio.
Does running Eigent from source need an account?
The preview path connects to Eigent cloud services and needs account registration. The README's Local Deployment path runs standalone with no cloud account required.
What do I do when a team outgrows a local desktop repo?
Kortix is the open-source AI Operating System: agents, skills, memory and connectors live in one git repo a team owns, each session runs on its own cloud computer, and work reaches main through a change request a person approves.
Run the repo, then run your company from a repo you own.
Self-host free, or start on Kortix Cloud.