The same platform, two front ends
Under the surface, both put static bytes at a URL on a global edge. One offers a page you drag a folder onto; the other offers an endpoint you call. The storage and the serving are the same class of thing; the interface is the whole difference.
Who each front end is for
| Author | Wants | Reach for |
|---|---|---|
| A person with a folder | A URL now, no config | A drop deploy |
| A program with a file | A verb it can call and re-call | The publish primitive |
Why an agent wants the endpoint
Because a loop cannot drag. It needs to publish, read the URL back, revise, and gate — all as calls it can make between turns. A page is something a human operates; an endpoint is something a program uses.
curl -X POST https://read.botook.ai/v1/publish -F file=@index.htmlThe privacy posture is not the same
A general deploy is usually public-by-default, because a person deploying a folder means to publish it. This primitive is unlisted-by-default, because a program assembling a page from its context did not necessarily mean to show it to the world. The default follows the author.
Running it yourself
The reason this exists as a repository rather than a service is that the platform's free tier is enough to host it on an account you own. No operator in the middle, no per-call charge — the infrastructure bills you directly.
Which to pick
If you are a human with a folder, the drop deploy is less friction. If you are wiring up an agent, the endpoint is the one that fits into a loop. They are not competitors so much as two answers to “who is publishing.”
Are these the same product?
No — one is a drag-and-drop deploy, the other is this publish primitive. They can share an edge platform because putting static bytes at a URL is the common part.
Does the primitive cost more to run than a drop deploy?
It runs within the platform's free tier on your own account, which is the entire reason it is a self-hosted repository and not a paid service.