ABOUT

Why this exists.

I spent twenty years building the networks that run Canadian enterprises. Routers, firewalls, MPLS VPNs, the gritty plumbing nobody notices until it breaks. In 2025 I started pivoting into AI — not as a theorist, not as a "prompt engineer," but as a builder.

The pivot came with a problem I couldn't solve by hiring someone: distribution. I wanted to build in public, but every night I was rewriting the same idea six times — for LinkedIn, for X, for a Reel, for Medium. The tools available were either dumb schedulers (still write N versions) or generic AI SaaS (generic output, not me). Neither scratched the itch.

Springy is what I built for myself, then cleaned up enough to give away. One idea goes in, fourteen platform-native drafts come out, in a voice that sounds like mine because it's been conditioned on my own past work.

What I believe about the shape of the problem.

LLMs are good at style transfer. They're not good at deciding what to post and when. The failure mode of "let an agent run your content" is generic output with no editorial point of view — the agent reverts to the mean.

The interesting problem isn't "can an AI write a LinkedIn post." It's: "can a system preserve a human's voice and judgment while scaling their distribution." Springy answers that by putting humans in two specific places (approve the anchor; approve the fan-out) and letting the machine do the translation work in between. The graph is rigid on purpose.

Why open source.

Three reasons:

1. The tool is the demo. Every install is a working example of what I'd build for a client. Closed-source would mean a landing page; open-source means a repo people can fork and judge.
2. I don't want to run a SaaS. I want to build more things. Springy needs to keep working without my attention — that means other people have to be able to fix it. MIT license + clear contribution path makes that possible.
3. I can't pretend to be a neutral vendor of a tool I use every day. If someone else builds a better cascade engine on top of this, good. I'd rather live inside a better ecosystem than gatekeep a worse one.

Who's behind it.

Anderson Andrade. Senior network architect (20+ years) pivoting into AI. Brazilian-Canadian, based in Aurora, Ontario. I also build OpenClaw (business automation), trading automation systems, and Adara Protocol (blockchain for autonomous AI agent economies — patent filed, named after my daughter).

You can find me on LinkedIn, X, and Bluesky. Those posts are, yes, cascading through a production springy install.

How to help.

• Star the repo. It costs you nothing and raises the signal when someone's deciding whether to try an alpha tool.
• Run the quickstart and tell me where it broke. Friction at step 2 tells me more than a feature request for step 15.
• Wire a platform I haven't covered. Most of them are ~150 lines. The plugin system was built for this.
• Discuss. Tell me what's confusing, what's missing, what you'd use it for that I didn't think of.

---

Springy is pre-1.0. Expect rough edges. It's released under MIT; you can fork it, sell consulting on top of it, or ignore me entirely. I'd prefer you file an issue first, but I'm not going to chase anyone down.