Who owns what we build

Not the same question as the disappearing vendor

The FAQ answers what happens if Optaimi stops existing: your agents run on infrastructure you own, so they carry on. That is about survival. This note is about the thing survival does not settle, which is whether the surviving deployment is a variant only we know, sitting on a branch nobody else can maintain. Those are different fears and they deserve separate answers.

Most of the work never reaches the core

Frameworks are built to be extended. Skills, plugins and MCP servers are the interfaces they offer for exactly that, and it is where the large majority of what we build for you lands, including most of the add-on menu. Work at an extension point leaves the framework itself untouched, so it adds nothing you have to carry when the framework moves on.

Sometimes it does reach the core, and we say so

We used to claim on this site that it never does. That claim was withdrawn on 19 August 2026 because it was not true. Some requirements can only be met by changing the framework itself. Making a framework accept a different authentication method so another tool can drive it is one. Teaching a framework a protocol it does not speak is another. Neither is exotic, and a promise that they will never come up is a promise made by somebody who has not done the work.

What we do instead is bound it. A change to the core is kept as a tracked patch against the upstream version you are pinned to, rather than as a rewritten copy of the project. It is listed in your handover document, so the difference between your deployment and stock upstream is a written list rather than an archaeology exercise. Where the change is general enough to be useful to anyone else, we put it upstream, and if it is accepted your deployment stops being different at all.

What a patch costs you

It is not free and it would be dishonest to present it as free. Every patch has to be reapplied and staging-tested against each upstream release, so a patched deployment carries more upgrade work than one that runs stock. On a retainer that work is ours. If you take the deployment in house, it becomes yours, and it is the main thing that makes a patched deployment harder to look after than a plain one. That cost is the reason we reach for an extension point first and the core last, rather than a reason to pretend the core is never touched.

Nobody here owns the orchestration layer

The four frameworks we deploy are all other people’s projects. None of them is ours and we control none of them, so there is no version of this where Optaimi is the gatekeeper of the layer your agents run on. Your rights to the framework itself come from that project’s own licence, direct from the people who publish it, named for your deployment at handover rather than granted by us. It is not ours to grant and we do not claim it.

What you hold

The deployment sits on infrastructure you own, and your configuration and your data are yours outright. What we build for you is licensed to you to run and to change, and that licence does not depend on you keeping a retainer, so it does not end when the engagement ends. The exact wording lives in the agreement signed before any work starts, because this is a page that publishes prices rather than terms, and a note is the wrong place to keep a contract.

The way out is a line on the menu

The strongest thing we can say here is not a reassurance. Framework migration is a published item on the add-on menu, priced on application because how long it takes is set by what you are moving. Moving you off a framework is work we sell rather than a conversation we avoid, which is the part a promise cannot fake.