Skip to main content

Lab Notebook · Minimalism — Pico Line

PicoAGI and PicoSaaS — the minimalism ideal, before the sprawl it didn't prevent

In January 2025 we built two deliberately minimal scaffolds in the same week: PicoAGI, a plugin-based Python AGI framework stripped to its essentials, and PicoSaaS, a lightweight SaaS starter with auth, sessions, and Stripe payments. Both were explicit exercises in 'strip to the minimum viable form' — a philosophy that shaped every later architecture, even though the sprawl that followed shows how hard that ideal is to hold onto.

JR
Jon RossFounder, vocabotics — 15 years building safety-critical systemsLab report · dated 20 January 2025
Verified by a human. Drafted with AI, verified by a human. Jon Ross, 20 Jan 2025
Living document. Reviewed 20 Jan 2025
Entry date
20 January 2025
Category
Agents
Access
🔓 Public
ResearchPartly proven

Two scaffolds, stripped to the minimum viable form, in the same week — the opposite instinct to the sprawl that, a year later, needed a cull to undo.

2
scaffolds: a minimal agent framework, a minimal SaaS starter
1 week
both built in the same January 2025 window
strip to minimum
the explicit design mandate for both

Honest evaluation

Partly proven

Both scaffolds worked exactly as designed and the minimalism instinct they carried genuinely propagated forward — but the report is honest that this ideal did not prevent the sprawl that followed, and needed a later, deliberate cull to correct.

What would prove or disprove it further

What would move this toward fully proven: evidence that projects built directly on these scaffolds stayed small over time rather than accreting complexity. What would move it toward disproven: if the instinct these scaffolds demonstrate turns out, on closer inspection, to have had no real causal effect on later architecture at all — the report currently asserts the influence but doesn't measure it.

The evidence — full reasoning behind the verdict

Verdict: partly-proven. As build artifacts, both scaffolds did what they set out to do: PicoAGI's plugin architecture let agents load and swap without touching the core, and PicoSaaS packaged the tedious parts of a SaaS starter into something small enough to read end to end. On that narrow claim — can you build a genuinely minimal, composable scaffold in each domain — the answer is yes.

The report doesn't stop there, though, and the more it says, the more mixed the picture gets. It names the durable win honestly (the "small and composable" instinct did carry forward into essentially every later architecture) but pairs it with an equally honest negative: the minimalism on display here was the ideal, and the portfolio that followed sprawled rather than staying small, requiring a deliberate cull sixteen months later to undo. PicoAGI and PicoSaaS did not, in the report's own words, prevent that sprawl.

Two small projects, built in the same week, both making the same argument: strip it to the minimum viable form. It's an ideal worth recording honestly — including how hard it turned out to be to keep.

What it was

PicoAGI: a minimal Python AGI framework with a plugin architecture for loading modular agents, driven by a rich CLI. PicoSaaS: a lightweight SaaS starter — auth, sessions, Stripe payments — built on Node.js, Express, Prisma, SQLite, Passport.js, and EJS. Different domains, same brief: build the smallest thing that actually works, and no more.

What we built

Research

PicoAGI's plugin architecture let agents be loaded and swapped without touching the framework core — composability over a monolith. PicoSaaS covered the genuinely tedious parts of a SaaS starter (authentication, session handling, payment integration) in a template small enough to actually read end to end, rather than a sprawling generator nobody fully understands.

Neither is an application on its own. Both are scaffolds — the point was never PicoAGI or PicoSaaS themselves, but the discipline of building something small enough to hold in your head.

What we learned — including the honest negative

The "small and composable" philosophy here carried forward into essentially every later platform and agent architecture in the portfolio — it's a real, durable design instinct, not a one-off.

But it's worth naming the tension honestly rather than only claiming the credit: the minimalism on display here was the ideal, and the sprawl that actually happened over the following year was closer to the opposite of it — dozens of projects, many later overlapping, chasing the capability curve. PicoAGI and PicoSaaS did not prevent that sprawl. They're kept in the record as evidence the instinct was there early, even when the follow-through wandered.

Where it went / status

Both remain scaffolds, not products — archived as reference designs rather than developed further as standalone tools. Their real legacy is architectural: the minimum-viable-form instinct they demonstrate resurfaces, more consequentially, in the Great Cull sixteen months later, when that same instinct had to be applied retroactively to everything that had sprawled in between.

What is still open — kept visible

The honest edges, next to the wins. This is what turns 🔬 into 🟢 — honestly.

  • Both are boilerplate/framework scaffolds, not applications — the value was in what they made possible later, not in themselves.
  • The minimalism here is a real ideal, but a genuine tension worth naming: the portfolio that followed sprawled rather than staying small, and needed a deliberate cull sixteen months later to correct course.

Where this connects

Sources

  1. vocabotics project audit — PicoAGI (minimal AGI framework) and PicoSaaS (minimal SaaS starter), Jan 2025vocabotics internal project history · as of January 2025

    We use cookies.