Skip to main content

How to Use Claude for ADHD: The Setup I Run Every Day

Published Chudi Nnorukam 10 min read

How to use Claude for ADHD without writing code: the brain dump, the one-question rule, the checkpoint, and the four ways it fails. The setup I run daily.

Why this matters

Claude has no ADHD mode. What works is using it as an external executive function: dump the whole overwhelming thing into it and ask for steps, ask one question per message, make it show evidence before you accept "done", and write a checkpoint note every 45 minutes so an interruption costs you thirty seconds instead of a rebuild. This page is the non-developer version of the setup I run every day. It also lists the four ways it fails.

In this cluster

Cluster context

This article sits inside Neurodivergent Systems.

Open topic hub

ADHD traits as engineering advantages: pattern recognition, resilience, and system design.

Most productivity advice assumes a brain that prefers linearity and uniform energy. This cluster is the counter-program: use neurodivergent traits as a design constraint and build systems around them.

Part of my Neurodivergent Systems series on tools and workflows for technical work.

Short answer: Claude has no ADHD mode. There is no setting to switch on. What works is a way of using it, and that way is the same whether you write code or not: hand Claude the job your executive function does badly, then keep it honest.

I have ADHD and I build software with Claude every day. Most of what I have written about it, including the ADHD engineer’s productivity system this page grew out of, is aimed at developers. This page is for everyone else. If you only ever use the Claude app or the website, everything below still applies. Nothing here needs a terminal.

If you are a developer and want the file-based version, read the CLAUDE.md guide for ADHD developers instead. This page stops before that layer.

What “using Claude for ADHD” actually means

My brain generates ideas faster than it can organize them. Thoughts arrive like popcorn: random, unpredictable, piling up. Traditional advice says capture them in an inbox. The inbox then becomes one more pile, and one more thing to feel guilty about.

Claude became my external prefrontal cortex. When a project overwhelms me, I dump the whole thing at it and ask for steps. When I cannot tell what matters, I describe the situation and ask for a priority order. When I am stuck, I ask what I am probably forgetting.

It does not judge. It does not get tired. It does not sigh when I bring the same problem back for the fifth time. It just processes.

That is the whole idea. The rest of this page is the mechanics: what to type, what rules to keep, and where it breaks.

Step 1: The brain dump

The first move is the one most people never make because it feels messy. Put everything in one message. Not a tidy request. Everything.

The mess is the point. Sorting is the executive-function task that ADHD makes expensive. Sorting is exactly what you are handing off.

Here is the shape I use:

I am overwhelmed by [the thing]. Here is everything in my head about it,
in no order:
- ...
- ...
- ...
Turn this into steps. Each step should take under one minute to start.
Tell me which step to do first, and only that step.

Two details in that prompt do the real work.

“Under one minute to start” is not the same as “under one minute to finish”. Task initiation is the wall. A step that takes an hour to finish but ten seconds to start is fine. A step that takes an hour to figure out how to begin is the enemy.

“Only that step” matters because a list of twelve steps is a new overwhelm. You wanted one door, not a hallway. Ask for one door.

Before I did this, breaking a project into pieces took me half an hour of staring, getting overwhelmed, and giving up. Now I describe the outcome, Claude breaks it into four or five pieces, and I pick one. I am no longer deciding “what do I do?” I am deciding “which of these four do I want first?” The second decision is a hundred times easier.

Step 2: One question per message

This is the rule I break most and regret most.

When three questions arrive in my head at once, the ADHD reflex is to fire all three into one message. What comes back is a blended answer that half-addresses each, and now I have to untangle it. Untangling is another executive-function task. I just created work for the part of my brain that was already struggling.

So: one question, one message. If you have three questions, send three messages. Context switching costs are real, and they are real for the tool as well as for you. A single clear question gets a single clear answer you can act on without parsing.

Built for this brain

ADHD Engineer's Productivity System $19

Notion templates built for how ADHD brains actually work: energy-based task management, two-axis filtering, and a brain-dump capture system.

Get the system

Step 3: Evidence before “done”

Claude will tell you something “should work”. Do not accept that.

This rule came from code, where “should work” means a bug you find at midnight. But it applies to everything. If Claude drafts an email and says it covers your three points, ask it to quote the sentence that covers each point. If it says a plan fits inside a week, ask it to lay the days out. If it summarizes a document, ask it for the passage each claim came from.

Decision paralysis dies when the evidence is clear. Vagueness feeds it. “This should be fine” leaves your brain with a loop to keep running. “Here is the line that does it” closes the loop.

The prompt is short:

Before you tell me this is done, show me the evidence.
Quote the exact part that does what I asked.

You will be surprised how often the evidence is not there.

Step 4: The checkpoint note

This is the habit that changed the most for me, and it is the one that has nothing to do with Claude being smart.

Every 45 minutes, I pause and write down three lines: what I was doing, what is next, and what is blocking me. In my developer setup that goes into a file Claude reads automatically. In the app, it goes at the top of the conversation, or into a Project’s instructions if you use those.

Why 45 minutes? Because that is roughly the size of a task I can hold in one piece. Work that fits in 45 minutes gets finished. Work that needs three hours gets abandoned at hour two when something shinier appears.

Why write it down at all? Because of interruptions.

The research on this is old and consistent. Gloria Mark’s work at the University of California, Irvine, found it takes about 23 minutes to get back to a task after an interruption. For a brain with ADHD, that number is the floor, not the average. Every interruption is a full rebuild: what was I doing, why, what did I already decide.

The checkpoint note turns a 23-minute rebuild into a thirty-second re-entry. “Working on the budget spreadsheet, stuck on the insurance line, next is the car column.” That sentence is the difference between an afternoon and a lost afternoon.

My one-minute rule: an interruption under a minute needs no note. An interruption over a minute does. Write the note before you deal with the interruption, not after. After is too late; the context is already gone.

Step 5: Ask what you are forgetting

The last regular prompt in my day is the simplest one.

Here is what I have done and what I plan to do next.
What am I probably forgetting?

ADHD working memory drops things silently. You do not notice the thing you forgot, because noticing it is the exact function that failed. A second processor with the whole context in front of it catches what you cannot.

This is not Claude being clever. It is Claude being outside your head. That is the entire advantage.

The four ways this fails

I have run this setup long enough to know exactly where it breaks. If you skip this section, you will find these out the expensive way.

Hyperfocus

Claude makes it very easy to go deeper than the task needed. You asked for a plan for the week and forty minutes later you are designing a system for tracking weekly plans. Novelty is the ADHD drug of choice, and a tool that can build anything is an endless supply of novelty.

Guard: the 45-minute checkpoint. When the timer fires, you re-read your own note and ask whether the thing you are doing now is the thing the note says. Usually it is not.

Interruption cascades

One interruption becomes three. A message arrives, you answer it, that raises a question, you look it up, the lookup reminds you of something else. Twenty minutes later the original task is a stranger.

Guard: the one-minute rule above. The note goes down before the second interruption can arrive.

Confident wrong answers

Claude is fluent when it is right and exactly as fluent when it is wrong. For an ADHD brain that is dangerous, because fluency feels like closure and closure is what you are craving. You accept the answer to make the open loop stop.

Guard: evidence before done. Every time. Especially when you are tired, because tired is when you most want the loop to close.

Treating the tool as the plan

The subtle one. Using Claude can feel productive in the way that buying a planner feels productive. You have set up the Project, written the instructions, had a great conversation about your goals. And done nothing.

Guard: the brain dump always ends with one step, and the step gets done before the next message. If you have had three conversations about a task and taken zero steps, the tool has become the avoidance.

What the setup costs

Two minutes at the start of a session to paste the checkpoint from last time. Thirty seconds every 45 minutes to update it. One extra message per task to ask for evidence. That is the whole overhead, and it is the overhead of a system that has held for me across bad weeks, not only good ones.

What it does not cost is the thing ADHD systems usually cost: setup. There is no app, no board, no color-coded calendar to abandon in week three. The system is five sentences you type and one note you keep. If you stop for a month and come back, the note is where you left it and the sentences still work.

What Claude does not do

It does not remember you between conversations unless you set that up. In the app, that means a Project with instructions, or pasting your checkpoint at the top. Assume it starts fresh and you will not be disappointed.

It does not know when you last did the task, whether you slept, or what your energy is like today. You tell it, or it plans as if every day is the same. Every day is not the same. I plan differently on a low-energy day, and I say so in the first message.

It does not replace medication, therapy, or a clinician. I am a developer describing a setup. The setup helps me get things done. It does not treat anything.

Where to go from here

If this page was enough, the next thing to do is not read more. Open Claude, paste the brain-dump prompt, and put one real thing into it. One.

If you want the deeper system this grew out of, the pillar is the ADHD engineer’s productivity system. It covers the energy-aware planning and the forgiveness rules that this page only touches.

If you write code, the same ideas become a file that Claude reads automatically every session, and the setup gets much stronger. Start with the Claude Code ADHD workflow, then the skills I run.

And if you want the prompts on this page as a set you can copy, with the failure each one guards, they are collected in ADHD prompts for Claude.

The whole setup, written up as a system you can install in an afternoon, is the ADHD Productivity System on the products page.

· Frequently asked

FAQ

Does Claude have an ADHD mode?

No. There is no toggle, setting, or official feature by that name from Anthropic. The phrase comes from a viral video. What people mean by it is a way of prompting and a set of habits, which this page describes.

Do I need to know how to code to use Claude for ADHD?

No. Everything on this page works in the Claude app or on the website. Developers get an extra layer from a memory file in their project, and I link to that guide, but the brain dump, the one-question rule, and the checkpoint need nothing but a text box.

Is this medical advice?

No. I am a developer with ADHD describing my own setup. Nothing here replaces a clinician, medication, or therapy. If you are deciding on treatment, talk to someone qualified.

Which is better for ADHD, Claude or ChatGPT?

I have used both as an external processor. The habits on this page do not depend on the model. I use Claude because I also build with it, so one tool holds my context across both kinds of work.

· Sources & further reading

Sources & Further Reading

Further reading

Reading Path

Continue the Neurodivergent Systems track

Go to hub

Built for this brain

ADHD Engineer's Productivity System $19

Notion templates built for how ADHD brains actually work: energy-based task management, two-axis filtering, and a brain-dump capture system.

Get the system

Want more of this in your Google results?

What do you think?

I post about this stuff on LinkedIn every day and the conversations there are great. If this post sparked a thought, I'd love to hear it.

Discuss on LinkedIn