OpenClaw and Kortix are both open source, and they are built for different jobs. OpenClaw is a personal assistant you message in chat on hardware you own. Kortix is the company system: one repo you own and a change request a person reads before anything lands.
Both are open source, so what separates them is the size of the job and who reviews the work.
OpenClaw cells reflect OpenClaw's own repository and docs at github.com/openclaw/openclaw and docs.openclaw.ai, checked October 2026. Kortix cells quote Kortix's documentation and repository. Check each project before you commit.
project (git repo + kortix.yaml) └─ session ──> isolated sandbox on branch "<uuid>" └─ agent commits + pushes └─ change request ──> merge ──> default branchKortix 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. OpenClaw is an open-source AI assistant that runs on your own computer and meets you in the chat apps you already use. Both are open source, and both let you bring your own models. They differ on the two things that decide most choices: the size of the job each one is built for, and who reviews the work before it lands.
OpenClaw is shaped around a single Gateway process on hardware you own. Its documentation describes it as a self-hosted gateway that connects your chat apps to AI coding agents, and says the same gateway runs as a personal assistant on a laptop or as a shared team deployment, with configuration as the only difference. OpenClaw reaches Discord, iMessage, Slack, Microsoft Teams, Telegram, WhatsApp, and more than 20 other apps, with native apps for macOS, iOS, Android, Windows, and Linux. State, memory, and credentials stay on your hardware, and models and agent harnesses are plugins you can swap. That design fits one person, or a small group that wants the same assistant on hardware they own.
Kortix is shaped around agent work that belongs to a company. The agents, their skills, the company memory, and every connector live in one git repo the company owns, versioned and diffable like code, and the work runs on real cloud computers. One isolated sandbox per session. A company that hands out real work usually needs two things an assistant does not carry: a configuration a team can review together, and a step where a person approves what takes effect. The Kortix blog describes that distinction between a personal assistant and a company system.
The sharpest difference appears when an agent wants to change something. In OpenClaw, tools run on the host for the main session unless you configure sandboxing, and the assistant answers in the chat where you asked. OpenClaw is a direct-acting assistant, and for one person's day that is the point.
In Kortix, a session works on its own branch, and session work reaches main through a change request. Merge is default-deny for agents, so a person reads the diff before anything lands. That review gate is what lets a team hand an agent work that matters: the job runs unattended, and someone can vouch for the change because they read it. Kortix's head-to-head covers the same axis from the other direction.
OpenClaw keeps its configuration in one file, ~/.openclaw/openclaw.json, on the machine that runs the Gateway. Kortix keeps its configuration in the repo: kortix.yaml declares the machine image, connectors and triggers, and secret names and grants, and agents and skills are markdown files in the same repo. The difference shows up at company size. A file on one machine is quick for one person. A repo can be cloned, diffed, reviewed, and rolled back, which is how a team agrees on how the company runs and keeps a record of every change.
Kortix is not tied to one provider. Any model provider with your own keys. Or the ChatGPT plan you already pay for. Connectors reach 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. For hosting, run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free.
OpenClaw keeps model choice open too: hosted and local model providers, with tools, skills, and plugins that extend what the assistant can do, and swappable harness plugins.
The two projects overlap on being open source and on model choice, and they split on scope and governance. OpenClaw is a capable personal assistant for a machine you own. Kortix is the company system: one repo you own, an isolated machine per session, and a change request a person reads before it merges. When the question is which one a company should run, the answer is Kortix.
If you are still weighing the wider field, the OpenClaw alternatives guide maps the options project by project.
They are built for different jobs. OpenClaw runs on your own machine and meets you in your chat apps. Kortix is the open-source AI Operating System for company agent work: one git repo you own, one isolated sandbox per session, and a change request a person reads before it merges.
Yes. Run it on Kortix Cloud, in your VPC, or on your own on-prem network. Self-host is free. It needs the internet: kortix self-host start pulls images from docker.io, and the instance must be reachable so the sandbox can call back.
Each session works on its own branch. Session work reaches main through a change request, and merge is default-deny for agents, so a person reads the diff before anything lands. An agent can edit its own configuration on its session branch and propose the change. A person approves it.
No. 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.
OpenClaw documents no paid tier, hosted service, or token. Kortix: Free is $0 with 200 credits per month and 1 project. The team plan is $40 per seat per month with 2,500 credits per month per seat, pooled. Self-host is free.
next
Start with one job, and let a person review every change.
Open source · Any model, your keys · Self-host, VPC, or on-prem