OpenHands pricing, and what the open-source Kortix costs

Is OpenHands free? The self-hosted build and the cloud

OpenHands is an open-source coding-agent platform from All Hands AI that runs autonomous agents against a codebase. OpenHands is free to self-host. Its own documentation describes a full open-source build made for a single user that runs on your own machine with your own model key, and the OpenHands repository grants the rights to use, copy, modify and redistribute that core. A team can run OpenHands inside its own infrastructure and pay the vendor nothing. What remains is the model behind the agent and the machine that hosts it, and the vendor's pricing page lists the local option as free and says self-hosters bring a key or serve an open-weight model themselves.

What OpenHands Cloud costs

OpenHands Cloud is the managed, hosted version of OpenHands that All Hands AI operates for you. As of October 2026 the vendor's pricing page lists three columns: open source, Individual and Enterprise. Open source and Individual both start free. The Individual tier allows one user and caps usage at 10 conversations a day, and you either bring your own model key or pay OpenHands's provider at cost with no markup. Organizations that need SAML single sign-on, centralized controls or a private VPC move to Enterprise, where the price is a custom quote. Model spend sits on top of every tier, because the agent's tokens are billed by the model provider whether you run in the cloud or on your own machine.

What Kortix costs, and why the shape differs

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. Kortix is free to self-host. The managed cloud is a $40/seat/month team plan from Kortix, which includes 2,500 pooled credits per seat, optional managed model access and seats for the people on your team. Agent Computer compute and managed model tokens draw from the same pool. A free tier includes 200 credits each month for sandbox compute and 1 project, and you can bring your own API key or connect your ChatGPT subscription instead of paying Kortix for models. Compute is billed per second, per resource, about $0.20/hour for the default 2 vCPU / 4 GiB / 20 GiB machine, and nothing while it is stopped.

Which cost shape fits a team whose work is wider than coding

OpenHands prices one lane, coding work. Its shape is free self-hosting, a free cloud entry tier, and a quote for the governed control plane. Kortix prices a workforce by seat: free self-hosting, $40/seat/month plus usage for a team, with coding as one job among several. If the work you want to delegate stays inside a codebase and your team already runs agent infrastructure, OpenHands's self-hosted path costs less. If the same agents also need to answer in Slack, run on a schedule and touch the apps the rest of the company uses, the real line item is the platform around the agent rather than the agent itself. Kortix runs coding and non-coding work on the same agents, from one repo that holds the agents, skills, memory and connectors, so a team does not buy a second platform for the non-engineering lanes. A seat price keeps the forecast steady as more people use it, because the bill follows people rather than conversations.

Compare the two directly in OpenHands vs Kortix, check whether the core is open source in is OpenHands open source, and see how to run your own in self-hosting.