I use coding tools to organize my life (and it works better than any productivity app)

I was in a 1:1 last week when someone on the team mentioned something important about a project. At the moment, I thought it was relevant, but I was focused on the conversation, following their reasoning. I didn’t stop to write it down.

Two days later, I needed to revisit that point. Normally, this would be a problem. But this time it wasn’t. I opened the meeting transcript, searched for the context, found exactly what they had said. I was able to pick up the thread and go deeper in the next meeting.

This completely changed my relationship with meetings. I no longer need to carry the cognitive load of “write down everything that seems important.” I can be present, truly listen, participate in the conversation.

I’ve tried to solve this before. Used Apple Notes for years, with the stylus. But my handwriting is terrible. And in the rush of the meeting, I’d write quickly in a way I couldn’t search for later. Or simply couldn’t read.

Until I had a strange insight: I use Claude Code every day to navigate massive code contexts. Thousands of files, architectural decisions from years ago, patterns that evolved through dozens of commits. And it works perfectly.

Why not use the same approach for meetings, decisions, and work contexts?

The problem nobody solved well

The real problem isn’t forgetting. It’s cognitive overload during meetings.

You’re in an important discussion, trying to follow the reasoning, contribute ideas, and at the same time need to decide: “is this important enough to write down?” And if so, how to write it quickly without losing the thread of the conversation?

I spent years using Apple Notes with the stylus during meetings. It seemed like the perfect solution: fast, natural, direct. But it had two fatal problems.

First: my handwriting is terrible. What I wrote hurriedly during meetings, I often couldn’t read later. Second: even when I could read it, I couldn’t find it. Searching in Apple Notes only works if you wrote legibly and used words you’ll remember later.

And when you need it most — that decision from three weeks ago, that comment someone made about a client, that insight that passed quickly in the conversation — it’s not there. Or it’s in an illegible note. Or it’s lost among dozens of other notes you never opened again.

I tried other things. Notion became a graveyard of organized pages I never revisited. Bullet journal required discipline I didn’t have. Obsidian consumed me organizing links between notes instead of paying attention to real work.

The problem isn’t lack of tools. It’s too much friction at the wrong time.

Why dev tools are different

Developers deal with massive context every day. An average codebase has thousands of files. Architectural decisions from years ago still impact new code. Dependencies intertwine in complex graphs that nobody can keep in their head.

And yet it works. Developers navigate this with ease.

The tools already solve the problem. Git records everything chronologically, without forcing you to categorize. Grep searches patterns in thousands of files in seconds. And AI, especially Claude Code, understands complete context without you explaining the structure.

The insight was realizing these capabilities work for any type of context, not just code. Meetings, decisions, projects, ideas — everything can be treated as “code” for storage and search purposes.

The fundamental difference: Claude Code is good at both things. Organizing AND searching.

It helps me organize when I need to. Suggests structures. Moves files. Reorganizes folders. But it’s minimal organization, sufficient, without excess.

And then, when I need to find something? The AI layer does semantic search. Understands context. Connects information that’s in different places.

Git records everything chronologically, without forcing me to categorize in elaborate systems. Claude Code navigates this history instantly. And when I ask something, it understands what I’m looking for, even if I don’t use the exact words.

Minimal organization + intelligent search. Both things together, with zero friction.

Which made me test this approach. And I discovered it works better than I imagined.

What I do in practice

I transcribe everything locally

Important meetings became transcriptions. I use Whisper running locally — doesn’t depend on cloud, no time limit, processes while I do other things.

A 1-hour meeting becomes a markdown file in 2 minutes. And most importantly: I can be 100% present in the conversation.

I no longer need to split attention between listening and writing. I don’t need to pause my thinking to decide if something is “important enough” to record. I don’t miss that quick comment someone made that later becomes crucial.

It’s all there. Searchable. Legible. Permanent.

I don’t transcribe everything. But any conversation that might matter later — 1:1s, strategic alignments, decision discussions — goes to text.

And when I need to revisit something? I talk to Claude Code normally: “What did person X say in the last meeting about project Y?” or “What do I need to do and can’t forget this week?”

Verbal context that used to vanish into thin air is now permanent. And I can finally be present in meetings.

Everything goes into a “brain” repo

I created a local Git repository called “brain”. Minimal structure: /meetings, /decisions, /projects, /ideas. Nothing elaborate. Nothing that needs maintenance.

brain/
├── meetings/
│   ├── 2026-02-10-quarterly-alignment.md
│   └── 2026-02-08-1on1-john.md
├── decisions/
│   └── 2026-01-15-stack-choice.md
├── projects/
│   └── cloud-migration.md
└── ideas/
    └── ai-experiments.md

Chronological commits with descriptive messages. Git tracks changes. Claude Code navigates all this like it navigates code.

What do I do when I need context? I talk normally to Claude Code inside the “brain” project: “When did we decide to migrate to Kubernetes?” or “What did we discuss in the last architecture meeting?”

Claude Code searches the files, connects information, responds. Instantly.

I don’t need to remember where I filed it. I don’t need to maintain structure. I just need to ask.

MCP as a bridge to other systems

Integration via MCP (Model Context Protocol) completely changed the game. Claude Code searches simultaneously in my local repo, in the company’s Notion, and in Linear where we track tasks.

I don’t need to leave the tool to access task context. I don’t need to open five tabs to understand a project’s status. Context comes to me, not me to it.

A concrete example: I need to prepare a presentation about quarterly progress. Before it was 30 minutes gathering information from Notion, Linear, Slack, documents.

Now? A question to Claude Code: “What was our progress on project X since January?”

It searches in three places, connects information, responds.

GitHub CLI to track work

Before any 1:1, I ask Claude Code: “What has John been working on the last two weeks?” It searches via GitHub CLI, lists PRs, commits, issues.

I don’t need to ask for updates. I don’t need the person to prepare a list of achievements. The work is already recorded, visible, accessible.

Recent commits reveal real progress. PRs show scope and complexity. Linked issues explain context. Everything that used to be “tell me what you did” is now “let me give you feedback on what I’ve already seen.”

Meetings become more productive. Less status update, more strategic discussion. Less “what” and more “why” and “how”.

Local repos as source of truth

Clone of the company’s main repositories locally. Claude Code analyzes commits, PRs, complete history. Technical context always accessible, even offline.

Someone mentions “that authentication refactoring”? Claude Code locates the PR, explains the decisions, shows the impact on related files. All in seconds.

Each commit tells a story. Claude Code reads the entire book in seconds.

What changed after 3 months

Three months using this daily. The changes are measurable.

I reduced time searching for context from 15-20 minutes to 2-3 minutes. Seems small, but it happens 5-6 times a day. That’s 60-90 minutes recovered, every day.

Decisions became more informed. Whenever I need to decide something, I search for complete context in minutes. I understand the history, review previous discussions, identify patterns.

Less anxiety about “forgetting something important”. It’s all recorded. It’s all searchable. If I need it, I find it.

Follow-up conversations became deeper. I review transcripts beforehand, arrive prepared with specific questions. No more “remind me what we talked about?”, but “about that point you raised…”

Concrete numbers so far:

• ~80 meeting transcriptions archived
• ~150 commits in the brain repo
• I search for context in Claude Code at least 5 times a day

And most importantly: I haven’t abandoned the system.

Because it doesn’t require perfection. Did I forget to transcribe a meeting? That’s okay. Didn’t commit an important decision? It’s probably somewhere else that Claude Code can access.

The system works even when I don’t work perfectly. And that’s rare in productivity systems.

It worked because it doesn’t require me to be perfect. Just that I use what I was already using.

Why this probably isn’t for you (and why it might be)

I’ll be honest: this isn’t for everyone.

It requires minimal comfort with terminal and dev tools. If you don’t use Git, don’t code, never ran a shell command in your life, the learning curve probably isn’t worth it.

If you don’t use AI and Claude Code daily, replicating this doesn’t make sense. The magic is in the tool that’s already in your workflow. Adding a new tool to “organize life” defeats the purpose.

But the concept can be adapted.

It doesn’t have to be Git. Could be Dropbox with advanced search. Could be Google Drive with date tagging. It doesn’t have to be Claude Code. Could be any LLM with semantic search capability in local files.

It doesn’t have to be local Whisper. Could be Otter.ai with export. Could be manual notes in plain text.

The principle is what matters: minimal organization + intelligent search works better than elaborate systems. Existing tools are better than new apps. Zero friction is more important than perfect system.

If you already use technical tools daily, adapt this to your flow. If you don’t, think about the concept: what do you already use today that has powerful search capability? How can you record more context in formats these tools understand?

Don’t copy what I do. Steal the principle.

The best tool was on the screen all along

Development tools already solved the problem of managing massive and complex context. Just adapt it to life.

If you already use dev tools, test for a week. Create a “brain” repo. Transcribe some meetings. Search like you search code. See if it works.

If you don’t, think about the principle: what do you already have on your screen every day that can search context better than any productivity app?

The best productivity tool sometimes isn’t in the productivity section. Sometimes it’s in the development section. And sometimes, it’s already open on your screen.

---

If this topic interests you, I’d love to exchange ideas. Find me on LinkedIn.