Open source and self-hosted, so you own the whole system

What open source and self-hosted change

Two properties decide how much of an automation system you actually control. Open source means the code that runs your work is code you can read, fork and audit. Self-hosted means the same system runs on hardware you control: your own VPC, your own on-prem network, or a managed cloud if you would rather not run it yourself. Kortix is the open-source AI Operating System. Read it, fork it, audit it.

A hosted workflow builder works the other way. You build in its editor, your steps and credentials sit in its database, and you reach everything through its dashboard. That fits a job that runs the same steps in the same order every time. It is a weaker fit for open-ended work, and it leaves a security review without an answer to where the configuration and the data rest.

One git repo is the whole system

Kortix keeps the parts as files in one git repo. Agents, skills, company memory, connector config and triggers are all text. You can grep the whole setup, diff any change, and roll any part of it back. Sessions run from that repo, so the file you edit is the system that runs.

An agent is who does the work. A skill is how the company does a job. Memory is what the system has learned so far. Each is a file, so a change to an agent or a skill reads as a diff you review. Per-resource permissions for people and agents. Roles, groups, and an audit trail. Secrets are encrypted at rest with a key per project. One isolated sandbox per session. Each session has its own isolated machine and branch. Thousands of agents in parallel on one config, each on its own cloud computer.

Work that lands as a change request

Because the configuration is text, the work can be reviewed like code. Session work reaches main through a change request. Merge is default-deny for agents. Approval gates you set. Off until you set them. An agent can edit its own configuration on its session branch and propose the change. A person approves it. The system improves one reviewed change at a time.

Where it differs from a hosted workflow builder

A hosted workflow builder rents you a fixed flowchart. It is quick to start and hard to leave: your flows, your credentials and the record of every run live in someone else's product, with nothing to fork or move. A self-hostable workflow engine such as n8n covers the hosting half of that question: you run the engine yourself and keep your flows on your own box. That settles where the data lives.

Kortix takes a different shape. Agents run on real cloud computers, reach your tools and hand back finished work. It reaches 3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP. Any model provider with your own keys. Or the ChatGPT plan you already pay for. The agents, the skills, the memory and the connectors stay in one repo you own. For the workflow-engine side of this choice, the Kortix blog compares the two models directly.

Start on your own infrastructure

Run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free. To start from source, install the CLI, scaffold the project and ship it, and the repo you pushed is the system that runs. The kortix self-host commands install and run the full stack on your own infrastructure. Connector credentials are brokered server-side and never enter the machine. Every action is recorded. Reading, exporting and streaming the audit log depend on the plan.

Compare it with the rest of the field on the alternatives page, or start from the home page.