For years now I’ve kept some of my social accounts alive with a Python script. It read from a collection of RSS feeds and shared at random times. It worked, my feeds stayed active, people asked me what was I doing reading at 3am.
The script had a bit of a personality: it logged out all the time, and tended to share links as simple urls, no preview. There was no context, no personal take, no small comment that explains why I was sharing this. Although I liked the serendipity of it, whenever I found a really cool article, I preferred to take the long route of curating the post by hand.
I tried Buffer and a few similar tools, but they felt extravagant for something I only do now and then. I don’t share articles for a living, I share them because I believe in elective affinities, that quiet collective pull toward certain ideas. A piece of writing isn’t fully alive until it’s passed from hand to hand. Unshared, it feels like the proverbial tree falling in a forest with no one around to hear it.
So I built EchoTree.
The goal
I wanted a light, intentional workflow:
- read the article in a clean view,
- pick a sentence or summary,
- write a short comment,
- and share now (or schedule).
That meant building:
- a reader mode preview,
- a comment box with a quick summary helper,
- a share‑now button,
- a queue for when I want to schedule.
How I built it
- PHP + Slim for routing and pages.
- Twig for templates.
- SQLite so it runs anywhere without setup.
- SimplePie + Readability for feeds and reader mode.
I used moltbot over WhatsApp while at the gym, which felt very 2026.
Codex 5.2 was behind the scenes. One strategy that worked extremely well was to draft together one Specs one Tasks document. Specs provided the business needs and Tasks supplied discrete steps to build the application.
Then I wrote “read the specs and implement the tasks one step at a time” and the thing did like 3 days of work in 5 minutes. I’d say the initial product was like 80% finished.
The whole thing lives in a folder and doesn’t need a database server or a queue.
You can find it at https://github.com/danielpradilla/echotree
