ChatGPT Work vs Kortix: an open-source system you own

What ChatGPT Work is

ChatGPT Work is OpenAI's agent for longer, multi-step tasks, launched on July 9, 2026. OpenAI describes it as "an agent in ChatGPT that helps you take on more ambitious tasks." It gathers context from the apps and files you connect, then produces finished materials such as sheets, slides, docs, and web apps, and it can stay on a project for hours by breaking the work into smaller steps and completing them on its own (launch announcement, product page).

OpenAI powers ChatGPT Work with its GPT models. It runs in OpenAI's cloud on web and mobile, or on your own computer through the desktop app (OpenAI Help). The configuration that shapes ChatGPT Work lives inside OpenAI's product: you connect tools through OpenAI's plugins, keep context in Projects and ChatGPT Space, and workspace admins set access, network rules, and spend controls in OpenAI's admin console.

What Kortix is

Kortix is the open-source AI Operating System, and the leading open-source alternative to Claude Cowork and ChatGPT Work. Your agents, their skills, your company memory, and every connector live in one git repo you own, and each session runs on its own cloud computer. Kortix is open source: read it, fork it, audit it, and the project carries 20,000+ GitHub stars (Kortix on GitHub).

Kortix is also model-agnostic. Point it at any model provider with your own keys, or use the ChatGPT plan you already pay for. Run it on Kortix Cloud, in your VPC, or on your own on-prem network, and self-host is free.

The difference that decides the rest

Both products finish work with agents, and the choice between them comes down to ownership. Three differences carry the rest.

Where the configuration lives. In Kortix, the agents, skills, and connector rules are files in a git repo you own. kortix.yaml declares the machine image and the connectors, an agent is a markdown file plus a block in kortix.yaml, and memory is files that accumulate. You can grep the whole company, diff any change, and roll any part of it back. In ChatGPT Work, the same settings live in OpenAI's product, where the schema belongs to OpenAI.

Whose models run the work. ChatGPT Work runs on OpenAI's own models, chosen by OpenAI. Kortix is model-agnostic: pick the model per agent, per session, or per message, and bring your own API key from any major provider. Your agents are not tied to one vendor's model roadmap.

Who can change the system. In ChatGPT Work, a workspace admin manages who has access and what the tools may do. In Kortix, per-resource permissions govern people and agents, with roles, groups, and an audit trail. Approval gates you set are off until you set them. An agent can edit its own configuration on its session branch and propose the change, and a person approves it: session work reaches main through a change request, and merge is default-deny for agents.

To run Kortix on your own infrastructure, the self-hosting guide covers the install. To see the wider field of open source ChatGPT Work alternatives, the comparison on this site covers the others. If you are already on ChatGPT Work and want the steps, the migration guide lays them out.