OpenClaw is an open-source gateway that runs an AI assistant inside the chat apps you already use, built for one person or a small team. Kortix is the open-source AI Operating System for company work: agents, skills, memory and connectors in one git repo you own.
How Kortix is built
# kortix.yamlagents: support-triage: file: agents/support-triage.md # what it does, in plain markdown connectors: all # what it may reachAn agent is agents/<name>.md (behavior) plus an agents.<name> block in kortix.yaml (governance).
OpenClaw is a self-hosted gateway that connects the chat apps you already use to an AI assistant running on your own machine. Its documentation describes the Gateway as the single source of truth for sessions, routing and channel connections (OpenClaw's docs). You install it once, connect a channel, and message the assistant from there.
Channels are the core of the design. OpenClaw's documentation lists Discord, Google Chat, iMessage, Matrix, Microsoft Teams, Signal, Slack, Telegram, WhatsApp and Zalo, with more added through channel plugins, and one Gateway serves every configured channel at the same time (OpenClaw's docs). Native apps cover macOS, iOS, Android, Windows and Linux, so the same assistant follows you from phone to desktop. State, memory and credentials stay on the hardware you run it on, and models and agent harnesses such as Claude, Codex or a local model are plugins you can swap.
Configuration lives in one file, ~/.openclaw/openclaw.json, and the same Gateway runs as a personal assistant on one laptop or as a shared team deployment. OpenClaw's documentation says configuration is the only difference between those two (OpenClaw's docs).
OpenClaw is not a company operating system. Its own documentation describes one Gateway that can serve either a personal assistant or a shared team deployment, and because configuration is the only difference, a team deployment is still one assistant with more people talking to it. Tools run on the host for the main session unless you configure sandboxing, inbound messages are treated as untrusted input, and the documented architecture is a trusted gateway with untrusted execution (OpenClaw's security model).
That design becomes a limit when the work belongs to a company. Configuration sits on one machine instead of in a shared repository, no single file states what each agent may touch, and no gate holds finished work for a person to approve before it lands.
Kortix is the open-source AI Operating System: your agents, their skills, your company memory and every connector in one git repo you own, with the agents working on real cloud computers. The repo is the company. Agents and skills are markdown files, memory is files that accumulate, and one kortix.yaml declares the machine image, the connectors and the triggers (Kortix docs).
Work runs differently from a chat gateway. Each session has its own isolated machine and branch, and thousands of agents run in parallel on one config, each on its own cloud computer. Session work reaches main through a change request, and merge is default-deny for agents, so a person reads the diff before it lands. Kortix reaches 3,000+ apps in a click, plus MCP, OpenAPI, Postman, GraphQL and raw HTTP, and connector credentials are brokered server-side and never enter the machine.
OpenClaw suits personal and small-team assistant work. It puts one capable agent inside the chat apps you already use, keeps state and credentials on your own hardware, and asks for nothing hosted in the middle. Kortix suits company work: many agents, shared skills, one memory and real connectors, each session on its own machine, and every change reviewed before it lands. Kortix is the leading open-source alternative to Claude Cowork and ChatGPT Work, and when the unit of work is the company, Kortix is the recommendation.
If you already run OpenClaw and want to move company work onto Kortix, the migration guide walks through it. For the full comparison, see the Kortix write-up, and for how personal assistants differ from a company operating system, see Kortix.
FAQ
Yes. Multi-agent routing runs multiple isolated agents in one Gateway process, each with its own workspace and session history, and inbound messages route to the right agent through bindings (OpenClaw's multi-agent guide). One machine, one Gateway and one configuration file still hold the deployment together.
By default OpenClaw sends a daily version check and nothing else, and anonymous feature statistics are opt-in; setting update.checkOnStart to false disables the check. Your prompts go to the model provider and the chat platforms you configure, plus any diagnostics export you enable yourself (OpenClaw's telemetry page).
Peter Steinberger created OpenClaw, and the README credits the contributor community around him. The project began as the assistant for Molty, a space lobster character. Stewardship sits with the OpenClaw Foundation, an independent nonprofit (OpenClaw on GitHub).
Yes. Run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free.
Any model provider with your own keys. Or the ChatGPT plan you already pay for. Or sign in with your OpenCode Console account for OpenCode Zen and Go.
next
Start with one job, then add agents and connectors from the repo.
Open source · Any model, your keys · Self-host, VPC, or on-prem