How to Use the P.A.R.A. Method on Your Phone — Without the Overhead

P.A.R.A. is one of those frameworks that makes perfect sense when you read about it and then quietly falls apart when you try to maintain it in real life. Not because the framework is wrong — it is genuinely useful — but because the filing step is where most people give up.

You capture a note. Now you have to decide: is this a Project, an Area, a Resource, or an Archive? That decision takes a few seconds. Multiply it by every note, every day, for months. The overhead accumulates. Eventually the inbox grows faster than you can sort it, and the system collapses under its own maintenance cost.

This post is about what happens when you remove that filing step and let AI handle it instead. We built this into Memex, so we are obviously partial. But the underlying problem — P.A.R.A. is great in theory and exhausting in practice — is real regardless of which tool you use.

A quick refresher on P.A.R.A.

Tiago Forte's P.A.R.A. method organizes all information into four categories:

• Projects — active efforts with a defined end date. A job application, a trip plan, a product launch.
• Areas — ongoing responsibilities without an end date. Health, finances, relationships, career development.
• Resources — topics you are interested in or learning about. Cooking techniques, programming languages, book notes.
• Archives — inactive items from the other three categories. Completed projects, past interests, old reference material.

The power of the framework is that it forces a decision about actionability. Instead of dumping everything into a flat list, you sort by how relevant something is to your current life. That makes retrieval faster and keeps your active workspace clean.

The weakness is that the sorting is manual. Every piece of information needs a human decision about where it belongs. On a desktop with Notion or Obsidian, that is manageable if you are disciplined. On a phone, where you are capturing thoughts between meetings or while walking, it is friction that kills the habit.

Why P.A.R.A. is harder on mobile

Most P.A.R.A. implementations assume a desktop workflow. You sit down, review your inbox, and file things into the right category. Notion databases, Obsidian folders, Tiago Forte's own courses — they all describe a deliberate sorting session.

But the capture happens on your phone. A thought while commuting. A photo of a whiteboard. A voice memo after a conversation. These fragments arrive at moments when you have no energy or context to decide whether something is a Project or a Resource.

The result is predictable: the inbox fills up, the sorting session gets postponed, and the system degrades into a flat dump of unsorted notes. The framework is still correct. The human maintenance cost is just too high for most people to sustain.

What if the filing happened automatically

This is the bet Memex makes. When you record something — text, a photo, a voice memo — a PKM agent analyzes the content and files it into the appropriate P.A.R.A. category. It creates Markdown files in a directory structure that mirrors Projects, Areas, Resources, and Archives. It links related records across categories. It updates the structure as new records arrive.

You do not see a filing dialog. You do not choose a category. You record, and the agent handles the rest. If it files something incorrectly, the knowledge base is editable — but the default path is zero-effort capture with AI-driven organization.

The agent also does something a manual system cannot: it re-evaluates over time. A Resource that becomes relevant to an active Project can be linked or reclassified. A completed Project moves to Archives. These transitions happen in the background as your records accumulate.

What the automated version loses

Honesty requires acknowledging what you give up. Manual P.A.R.A. filing forces you to think about each piece of information. That thinking has value. It is a form of review that helps you internalize what you have captured.

When AI does the filing, you lose that forced reflection. The system is more efficient but less deliberate. Some P.A.R.A. practitioners would argue that the filing decision itself is part of the learning process, and they would not be wrong.

The question is whether you actually do the filing. If you maintain a disciplined weekly review and enjoy the sorting process, manual P.A.R.A. in Obsidian or Notion is the better path. If your inbox has three hundred unsorted items and you have not reviewed it in a month, automated filing is a pragmatic improvement over no filing at all.

How it works in practice

A concrete example. You are walking home and voice-record: "Need to book flights for the Tokyo trip next month. Also, that ramen place Jun mentioned — look it up later."

Memex transcribes the voice note using on-device speech recognition. The Card Agent generates two cards: a task card for booking flights and a place card for the ramen restaurant. The PKM Agent files the flight booking under your Tokyo trip Project and the restaurant under a Resources topic for Tokyo dining. If you later record a photo at the restaurant, the Insight Agent can link it back to the original mention.

None of this required you to open a filing dialog, choose a category, or even think about P.A.R.A. The framework is running underneath, but the interaction is just: record and move on.

This is not the only way to do P.A.R.A.

Memex's approach is one implementation. If you prefer manual control:

• Obsidian with a folder structure and the Daily Notes plugin gives you full manual P.A.R.A. with plain Markdown files.
• Notion with a P.A.R.A. template database gives you a visual, filterable system with cloud sync.
• Apple Notes with folders is the simplest version — no AI, no automation, just four folders.

Each approach trades off differently between control, automation, and maintenance cost. Memex sits at the automation end of that spectrum. If that is what you need, it is worth trying. If you value the manual sorting process, stick with a tool that preserves it.

Getting started

If you want to try automated P.A.R.A. on your phone:

• Download Memex from the App Store or Google Play.
• Connect a model provider (Gemini, OpenAI, Claude, or any of the twelve supported providers).
• Start recording. Text, photos, voice — whatever is natural.
• Check the Knowledge tab after a few days. Your records will be organized into P.A.R.A. categories as interconnected Markdown files.

The source code is on GitHub if you want to see how the PKM agent works under the hood. For more context on the product, read why we built Memex or our comparison of AI journal apps.

FAQ