Devs: It sucks ass.
We would have a senior dev on the dealing-making-or-breaking call and he would say something like, "Wait -- sorry, do you support <specific thing that our system doesn't have>?" We would say honestly that we don't have that yet. Then the senior dev would say, "Ooohh.... sorry... yeah... we need (per-message arbitrary data storage | support for on-prem | real-time analytics | per-messsage privacy controls | etc.)." We eventually had to face the reality that one of two things must be true:
1) The market is just too fragmented such that there cannot exist one product that serves it
2) The devs involved really, really don't want to give a product like this a chance because they actually want to build fun features like this themselves and they're (rightfully) skeptical that some random startup can do it better
As I sit here now, I have some confidence that it's 75% #2 and 25% #1. Anyway, the basic shape of the sales calls went a lot like what you're saying.
Due to the software culture I've been exposed to, especially in web-dev, I am cautious about any level of indirection added by third parties. They may seem like a value-add to first-order, but when managed in the context of a larger system, tend to make performance worse, debugging more difficult, and modifications high-inertia. Not as a principle, but based on observations.
For example, When I see text like this: > With pre-built components for chat, presence, and notifications, as well as fully customizable UI primitives, Cord enables rapid implementation of sophisticated, in-app collaboration tools.
my pattern-matching brain thinks This is probably going to be slow, require the user to download a large amount of data, and may or may not reduce code required to implement, and will make customization and modification tedius. Maybe this isn't true for this particular library, but it seems to be a good rule, currently.