A serverless implementation of the HTMX trello clone developed by rajasegar.
https://github.com/rajasegar/htmx-trello
https://trello-clone.jeremiahlangner.workers.dev
This is a simple application to implement with enough complexity to learn how HTMX interacts while flexing some stiff serverless muscles in a slightly different serverless paradigm than my daily driver (AWS). This particular architecture/stack is not necessarily recommended.
rajasegar's Heroku instance of this server example for HTMX with node.js is down. I don't believe HTMX needs a server so much as "state storage" and "content delivery" mechanisms (and in almost all cases, for HTMX applications, a server is more beneficial for both the developer and the user than a serverless implementation). That is, the backend should represent the state. In this case, I'm using Cloudflare's KV storage and may switch to Durable Objects for consistency's sake.
HTMX over serverless requires several implementations that are neither performant nor advantageous (for the individual/company developing and administering the application; the user will notice little difference). Because HTMX utilizes route/method based server access to update the client state it incurs a large request penalty on the serverless backend.
As with this application, I recommend using a service worker as a backend to serve the state and requests for HTMX.
For development, clone the repository, navigate to the repo locally, run
npm install, update the wrangler.toml file with your values, and run
wrangler dev.
To try it out, navigate to the app in your browser at https://trello-clone.jeremiahlangner.workers.dev.
The application loads a service worker and can function offline until any updates are necessary.
To save your board data to the remote database click "save" in the upper-right corner.
To download the data from the remote database click "sync" in the upper right corner.
This project is developed from code written by Rajasegar Chandran. He is included as a copyright holder in the license. This is not a direct fork, but most of his code is included in the worker scripts.