Become a sponsor to Gentleman-Programming
Hey, I'm Alan Buscaglia ๐
App Lead at Prowler, Angular Google Developer Expert, Microsoft MVP, speaker, and the creator of Gentleman Programming. But here's what I actually spend my time on, and why you might want to back me.
AI can execute insanely fast. But without structure, it also destroys context, mixes responsibilities, bloats PRs, invents decisions, and leaves you with code that "works" but nobody can maintain. Most people use AI as autocomplete on steroids. I think that's a trap, so I build open source to fix it.
๐ ๏ธ Gentle-AI โ a discipline layer for AI coding agents
Gentle-AI isn't another chatbot or a pile of loose prompts. It's a controlled development harness (built on Pi) that defines how an agent should think: when to ask, when to delegate, when to run Spec-Driven Development / OpenSpec, how to protect the human doing the review, and how to stop AI from becoming a technical-debt machine.
It makes your agent work like a small but disciplined team: a parent orchestrator holds the main thread, specialized subagents explore / implement / review / verify, SDD saves every decision as an artifact, delegation rules stop one agent from reading half the repo and editing blind, review-workload protection looks after the human, and Strict TDD forces real evidence (RED, GREEN, REFACTOR).
Instead of "do this and let's see," it pushes a real process: understand โ decide โ specify โ implement โ verify โ review โ deliver. That's engineering. The alternative is roulette with pretty syntax.
๐ง Engram โ persistent memory for your coding agents
Your agent forgets everything the moment the session ends or the context gets compacted. Engram gives it an external brain: it stores decisions, resolved bugs, project patterns, discoveries and useful context so any agent can pull it back later.
Under the hood it's a local-first Go binary built on SQLite + FTS5, reachable via CLI, MCP, HTTP API and TUI, and it works across agents (Pi, Claude, Codex, Gemini, Cursor and more). That turns your agent from something that "remembers this chat" into something that accumulates real project knowledge: no re-explaining context every session, decisions and conventions that survive compactions, and different agents sharing the same brain. And when you need it, Engram Cloud is the optional layer to sync, replicate and visualize that memory across machines or a whole team, without ever giving up local-first: your local SQLite stays the source of truth, sync is explicit, and you can even self-host it.
Both are free, open source (MIT), and already used by thousands of devs. Around them there's a whole community: +120,000 on YouTube, 15,000 on Twitch, 50,000 on Instagram, +10,000 in Discord, and +65 open-source repos. Mentorship, real projects, no gatekeeping, no hype.
So what does your sponsorship actually do? It buys me the time to keep building and maintaining Gentle-AI, Engram and the rest of the stack, free and open for everyone. Because the future isn't won by whoever "uses AI." It's won by whoever knows how to direct it with real criteria.
If that's the kind of engineering you believe in, come hang out. See you there ๐
53 sponsors have funded Gentleman-Programmingโs work.