Do I need a developer to set transglot up?

For the first connection, usually yes, but only once, and there are routes that need no code at all.

Anyone deciding how to get started.

There are three ways to get strings into transglot, and they need very different amounts of engineering:

No code. Import a whole project from another localization tool: Lokalise, Phrase or Crowdin. That is a screen in transglot; you need a read token from the tool you are leaving and nothing else.

Connect a system you already run. A code repository, a content platform, or a file-storage connector is set up once from the integrations screens. Someone with access to the other system has to authorise the connection, but that is configuration, not programming.

Push from your codebase. The command line tool, the API and the editor plugins are the route most engineering teams take, because it fits into an existing build. This is the one that needs a developer, and it is what the Quickstart snippets on a project's Get started page are for.

Getting translations back out has the same split: the CDN needs a few lines in your app, while a repository connection can open a pull request against your own files with no runtime change at all.