Method Guide

The StripFlowing method:
a FIFO protocol for human attention.

Your agents can run in parallel. Your attention cannot. StripFlowing gives the human operator a simple queue, a clean resumption point, and a repeatable rule for deciding what to touch next.

This page explains the method. The app at the root domain is where you actually run it.

Use this page to understand the system. Use the app to execute it.

Your brain has one execution lane. Your AI work does not.

The failure mode is familiar: you launch one AI run, another finishes first, a third needs input, you jump tabs to inspect it, then spend the next ten minutes reconstructing where you were in the original task.

The real bottleneck is not model throughput. It is the human supervisor repeatedly losing context. StripFlowing is a way to externalize that context so you stop paying the same re-orientation cost over and over.

"Everyone's building frameworks to orchestrate agents. Nobody's building a framework to orchestrate the human supervising them."

Working memory is narrow, and context switches are expensive. Once multiple runs are in flight, what matters is not reacting fast to every notification. What matters is having a stable rule for what to resume next and enough saved state to resume it without rummaging through tabs and chat logs.

Rule 1

Every delegated task gets a strip with only three fields: project, task, and anchor.

Rule 2

Every strip enters a queue. You do not pick the next task by mood, urgency theater, or notification order.

Rule 3

When you re-enter work, the anchor tells you exactly where to land, so resumption is immediate instead of reconstructive.


How it works

Three fields on a strip. A FIFO queue. One active review at a time. That is the whole operating system.

1

Write the strip

Every time you delegate a task to an AI agent, create one strip. Three fields only: Project (which context), Task (what the agent is doing), and Anchor (the exact place to return to, such as a chat, tab, branch, or tool window). The Anchor removes the search step from task switching.

2

Queue it in the bay

Put the strip into a visible queue. The newest strip goes on top. You always work from the bottom: the oldest task first, regardless of which agent finished first. This is FIFO for human attention. No chasing notifications. No reacting to whichever run happens to finish first.

3

Pull and advance

When you're ready, pull the bottom strip, look at the Anchor, and land exactly where you need to be — no search, no reconstruction, instant re-entry. Review the output, decide the next move, and if the task continues, put the strip back on top. If the task is done, remove it. The queue becomes your single external source of truth.

⚑
Cocking

Borrowed from Air Traffic Control: shift a strip sideways in the rack to signal a blocker, a critical error, or something that needs immediate attention. Unlike a notification that disappears, a cocked strip stays visually out of alignment until you handle it.


Why it works — the science behind it

StripFlowing is a practical synthesis, not a lab-validated framework. The logic behind it is still grounded in known constraints of attention, resumption, and external memory.

Cognitive Offloading

Writing a strip and placing it in the bay is a brain dump. You save state outside your head so your working memory can focus on one thing (Risko & Gilbert, 2016).

Attention Residue

Sophie Leroy (2009) showed that switching tasks without an explicit resumption plan leaves attention fragments behind. The strip is that plan — a Ready-to-Resume signal condensed into three fields.

Physical Writing Advantage

Writing on paper produces 25% faster task completion than digital equivalents, with significantly higher hippocampal activation (Umejima et al., University of Tokyo, 2021).

ATC Parallel

Air traffic controllers have managed dozens of aircraft in parallel for decades using physical Flight Progress Strips in a rack — even with digital radar available. Strips outperform dashboards for situational awareness (Mackay, 1999).


The physical setup

Start with any scrap of paper or a flashcard — that's all you need. If you want something sturdier, two 3D-printable models are available on MakerWorld.

Run it now Open StripFlowing and use the queue directly in the app. →
→