Prompt Library

Navigate Your Inner World with AI-Guided Reflection

45 copy-paste prompts

45 thoughtful ChatGPT prompts for therapy exercises, emotional processing, CBT techniques, and self-reflection: designed as a complement to professional care.

In short: This page contains 45 copy-paste ready prompts, organized into 9 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Drafted with AI writing assistance and checked against our editorial policy.

Louis Corneloup
By Louis Corneloup Β· Founder, Techpresso
Last updated Β·Drafted with AI assistance (how we make prompts)

Cognitive Behavioral Therapy (CBT)

5 prompts

Thought Record Exercise

1/45

✨ What it does

ChatGPT walks you through a thought record for [describe situation], starting from [emotion] at intensity [1-10] and your [automatic thought]. You answer only the questions you want, and you decide what you share.

Help me complete a CBT thought record. The situation that triggered my distress was: [describe situation]. The emotion I felt was [emotion] at an intensity of [1-10]. The automatic thought that went through my mind was: "[automatic thought]." Now walk me through these steps: (1) What is the evidence FOR this thought? (2) What is the evidence AGAINST it? (3) What cognitive distortion might this reflect (catastrophizing, mind-reading, all-or-nothing thinking, etc.)? (4) What is a more balanced alternative thought? (5) After considering the alternative, re-rate the emotion intensity.

πŸ’‘

Pro tip: This is not a replacement for therapy. Use it between sessions to practice skills your therapist has taught you.

Identify Cognitive Distortions

2/45

✨ What it does

ChatGPT maps each of [thought 1] through [thought 5] to a common thinking trap, such as all-or-nothing or jumping to conclusions. You read the labels, keep the ones that fit, and ignore the rest.

Here are some thoughts I have been having lately: 1. [thought 1] 2. [thought 2] 3. [thought 3] 4. [thought 4] 5. [thought 5] For each thought, identify which cognitive distortion(s) it represents from this list: all-or-nothing thinking, overgeneralization, mental filter, disqualifying the positive, jumping to conclusions (mind-reading or fortune-telling), magnification/minimization, emotional reasoning, should statements, labeling, personalization. Explain why it fits that distortion and suggest a more balanced reframe.

πŸ’‘

Pro tip: The goal is not to think positively but to think accurately. A balanced thought acknowledges difficulty while removing distortion.

Behavioral Activation Plan

3/45

✨ What it does

ChatGPT drafts a small activity plan for when you feel [depressed/low/unmotivated], based on [describe typical day] and [list activities] you stopped. You pick one easy step, and you skip anything that feels like too much.

I have been feeling [depressed/low/unmotivated] and I have stopped doing activities I used to enjoy. Help me create a behavioral activation plan. My current daily routine looks like: [describe typical day]. Activities I used to enjoy but have stopped: [list activities]. My energy level on a typical day is [1-10]. Create a gradual 2-week plan that reintroduces pleasurable and mastery activities, starting very small (things I can do even at low energy) and building up. For each activity, rate the expected difficulty (1-10) and expected pleasure (1-10). Include a mix of social, physical, creative, and routine activities.

πŸ’‘

Pro tip: Start with the easiest activities first. The goal is momentum, not perfection. Even a 5-minute walk counts.

Worry Time Script

4/45

✨ What it does

ChatGPT sorts [List 5-8 current worries] about [general topic] into problems you can act on and worries you cannot control, then suggests a daily worry window. You try the window, and you decide which items belong there.

I tend to worry throughout the day about [general topic]. Help me implement a "worry time" exercise. First, take these worries I have been carrying: [List 5-8 current worries] For each worry, categorize it as: (1) a practical problem I can act on, or (2) a hypothetical worry I cannot control. For practical problems, suggest one concrete next step I can take this week. For hypothetical worries, write a brief acceptance statement I can read when the worry surfaces outside of worry time. Then give me a script for what to tell myself when worries pop up during the day: how to acknowledge them and defer them to my scheduled worry time.

πŸ’‘

Pro tip: Schedule worry time for the same 20-minute window each day, ideally not before bed. Write worries down to externalize them.

Core Belief Examination

5/45

✨ What it does

ChatGPT walks a downward-arrow set of questions from a recent [describe situation] toward a core belief you name, such as not being good enough. You stop whenever you want, and you decide what you share.

I think I have a core belief that "[core belief, e.g., I am not good enough / I am unlovable / The world is dangerous / I must be perfect]." Help me examine this belief using the downward arrow technique. Start from a recent situation where this belief was activated: [describe situation]. Then: (1) trace backward: what experiences in my life might have formed this belief? (2) list evidence from my life that contradicts this belief, (3) identify how this belief affects my behavior and relationships, (4) suggest a more nuanced alternative core belief, and (5) give me 3 small behavioral experiments I could try this week to test whether the old belief or the new belief is more accurate.

πŸ’‘

Pro tip: Core beliefs often formed in childhood. Changing them takes time and repetition. Consider sharing what you discover with your therapist.

Emotional Processing

5 prompts

Name and Validate Emotions

6/45

✨ What it does

ChatGPT suggests more precise names for what you feel from [describe physical sensations, thoughts, and feelings as best you can], then reflects them back without judging. You keep the words that fit, and you leave the rest.

I am experiencing a complex emotional state right now. Here is what is happening in my body and mind: [describe physical sensations, thoughts, and feelings as best you can]. Help me: (1) identify and name the specific emotions I might be experiencing (use a nuanced emotion vocabulary, not just "sad" or "angry"), (2) validate each emotion by explaining why it makes sense given my situation, (3) identify if any emotions are in conflict with each other (this is normal), and (4) suggest a way to sit with these feelings without trying to fix or suppress them.

πŸ’‘

Pro tip: Research shows that people who can name emotions with specificity (e.g., "disappointed" vs "sad") regulate them more effectively.

Unsent Letter Exercise

7/45

✨ What it does

ChatGPT drafts an unsent letter to [person or situation] from [describe what happened and how it affected you], in the tone you ask for. You edit or scrap it, and you decide whether it stays private.

I have unresolved feelings toward [person or situation]. Help me write an unsent letter to process these emotions. The situation: [describe what happened and how it affected you]. I want to express: [anger/hurt/disappointment/grief/gratitude/forgiveness: or a mix]. Write a raw, honest letter from my perspective that: (1) describes what happened and its impact on me, (2) names the emotions I have been carrying, (3) says the things I never got to say, (4) acknowledges any complexity or mixed feelings, and (5) ends with what I want to release or carry forward. This letter will never be sent: it is purely for my processing.

πŸ’‘

Pro tip: After reading the letter, notice what resonates and what does not. Edit it in your own words. The processing happens in the writing, not the sending.

Grief Processing Guide

8/45

✨ What it does

ChatGPT offers gentle reflections on a loss you name, from [timeframe] and [describe current emotional state], without forcing stages or a timeline. You decide how much detail to share, and you stop when you need to.

I am grieving [describe the loss: person, relationship, identity, life stage, pet, etc.]. The loss happened [timeframe]. Right now I am feeling: [describe current emotional state]. I need help processing this grief. Please: (1) normalize what I am feeling by explaining where it might fall in the grief process (without rigidly applying stages), (2) identify any feelings I might be suppressing or judging myself for, (3) suggest 3 gentle grief rituals or practices I could try this week, (4) help me write a brief message to what I have lost: a way of honoring it, and (5) remind me what healthy grief looks like versus when I might want additional support.

πŸ’‘

Pro tip: Grief is not linear. You may revisit intense emotions months or years later. This is normal, not a setback.

Inner Child Dialogue

9/45

✨ What it does

ChatGPT drafts a dialogue between your adult self and a younger part, using [describe the experience or pattern] and [describe current behavior or emotional reaction]. You read it, change any line that feels wrong, and you decide what you share.

I want to practice an inner child exercise. A wound from my childhood that still affects me is: [describe the experience or pattern]. When this wound is activated today, I tend to [describe current behavior or emotional reaction]. Guide me through a dialogue between my adult self and my inner child. Have my adult self: (1) ask the child what they needed back then, (2) validate what the child experienced, (3) say the words the child needed to hear from an adult, and (4) make a specific promise about how the adult self will protect and care for the child going forward. Write this as an actual dialogue I can read aloud.

πŸ’‘

Pro tip: This can bring up strong emotions. Do this exercise in a safe, private space and be gentle with yourself afterward.

Emotion Tracking Analysis

10/45

✨ What it does

ChatGPT reads your week of [For each day, list: situation, emotion(s), intensity 1-10, and what you did] and names patterns in feelings, triggers, and how you coped. You keep the insights that ring true, and you ignore the rest.

Here is my emotion log from the past week: [For each day, list: situation, emotion(s), intensity 1-10, and what you did] Analyze my patterns: (1) Which emotions came up most frequently? (2) What situations or times of day triggered the strongest reactions? (3) Are there any patterns in how I coped (avoidance, rumination, healthy coping, etc.)? (4) What does this data suggest about my current emotional needs? (5) Based on these patterns, suggest 2-3 specific skills or practices I should focus on this coming week.

πŸ’‘

Pro tip: Track for at least 2 weeks to see meaningful patterns. Use a simple notes app or paper: do not over-complicate the tracking.

Anxiety Management

5 prompts

Grounding Exercise Script

11/45

✨ What it does

ChatGPT writes a grounding sequence for anxiety at [1-10], with counted breaths and a 5-4-3-2-1 sensory pass based on [describe anxious thoughts]. You follow only the steps that help, and you stop if it feels worse.

I am feeling anxious right now. My anxiety level is [1-10]. The thoughts driving it are: [describe anxious thoughts]. Walk me through a comprehensive grounding exercise. Include: (1) a breathing technique with specific counts, (2) a 5-4-3-2-1 sensory grounding exercise customized to my current environment [describe where you are], (3) a body scan to identify and release physical tension, (4) a brief reality-testing question to check if my anxious predictions are facts or fears, and (5) a compassionate closing statement I can say to myself.

πŸ’‘

Pro tip: Practice grounding when you are calm so it becomes automatic when you are anxious. The body responds faster to familiar techniques.

Anxiety Exposure Hierarchy

12/45

✨ What it does

ChatGPT builds a 10-step exposure ladder from [what I currently avoid] toward [ultimate goal] for anxiety around a fear you name, such as speaking or flying. You choose the next tiny step, and you never have to jump ahead.

I have anxiety around [specific fear or situation, e.g., public speaking, social situations, health concerns, flying]. Help me build a gradual exposure hierarchy. My current avoidance level: [what I currently avoid]. The goal I want to work toward: [ultimate goal]. Create a ladder of 10 steps from least anxiety-provoking (maybe 2/10 anxiety) to most anxiety-provoking (8-9/10 anxiety). For each step: describe the specific exposure, rate the expected anxiety (1-10), suggest how long I should stay in the situation, and provide a coping thought I can use during the exposure.

πŸ’‘

Pro tip: Never skip steps. Master each level (anxiety drops to 3/10 or below) before moving up. Exposure therapy works through repetition, not force.

Catastrophe Deconstruction

13/45

✨ What it does

ChatGPT walks questions that test [describe feared scenario] against a worst case, a best case, and a likely case, starting from [worst case scenario]. You answer only what you want, and you decide what feels true.

I am catastrophizing about [describe feared scenario]. My mind is telling me that [worst case scenario]. Help me deconstruct this catastrophe: (1) What is the absolute worst case? (2) What is the absolute best case? (3) What is the most realistic, probable case? (4) If the worst case actually happened, what would I do? Walk me through a concrete coping plan. (5) What have I survived in the past that I once thought I could not handle? (6) What would I tell a friend who had this exact fear? Write the response as a calm, grounded friend would: direct but compassionate.

πŸ’‘

Pro tip: The most powerful question is often "what would I actually do if it happened?" because it proves you have more resilience than anxiety credits you for.

Social Anxiety Preparation

14/45

✨ What it does

ChatGPT drafts prep for [describe event, who will be there, what is expected], including kinder thoughts, in-the-moment ideas, and aftercare. You use the pieces that help, and you skip anything that feels like too much.

I have a social situation coming up: [describe event, who will be there, what is expected]. My anxiety about it: [describe specific fears: judgment, awkwardness, not knowing what to say, etc.]. Help me prepare: (1) identify the cognitive distortions in my fears (mind-reading, fortune-telling, spotlight effect), (2) generate 5 open-ended questions I can ask people to take pressure off myself, (3) create an exit strategy I can use if I feel overwhelmed (without full avoidance), (4) write a pre-event pep talk that is realistic, not toxic-positive, and (5) design a post-event self-compassion plan regardless of how it goes.

πŸ’‘

Pro tip: People are far less focused on you than your anxiety suggests. This is called the spotlight effect: you are not under a microscope.

Sleep Anxiety Wind-Down

15/45

✨ What it does

ChatGPT writes a bedtime wind-down for racing thoughts like [describe typical bedtime anxious thoughts], including a worry dump and a body relaxation you can follow. You try the pieces that fit [describe], and you stop if sleep gets worse.

I struggle with anxiety at bedtime. When I lie down, my mind starts racing with thoughts like: [describe typical bedtime anxious thoughts]. My current bedtime routine: [describe]. Create a comprehensive sleep anxiety protocol: (1) a worry dump exercise I can do 2 hours before bed to externalize my thoughts, (2) a progressive muscle relaxation script customized for lying in bed, (3) a cognitive shuffling technique to break the anxiety thought loop, (4) rules for what to do if I have been lying awake for more than 20 minutes, and (5) a mantra or brief meditation script I can use when anxious thoughts return.

πŸ’‘

Pro tip: The bed should only be associated with sleep. If you are anxious for more than 20 minutes, get up and do something boring until sleepy.

Like these prompts? There are full tutorials behind them.

Learn the workflows, not just the prompts. Easy-to-follow tutorials inside AI Academy, and the library keeps growing.

Try AI Academy Free

Self-Compassion & Acceptance

5 prompts

Self-Compassion Letter

16/45

✨ What it does

ChatGPT drafts a warm letter to you about [describe situation where you feel shame, guilt, or self-criticism], answering the voice that says [what you tell yourself]. You read it in private, keep the lines that land, and leave the rest.

I am being very hard on myself about [describe situation where you feel shame, guilt, or self-criticism]. The critical voice in my head is saying: "[what you tell yourself]." Write a letter to me from the perspective of an infinitely compassionate friend who knows my full story. This friend should: (1) acknowledge my pain without minimizing it, (2) remind me that suffering and imperfection are part of being human, (3) point out what they see in me that I cannot see right now, (4) gently challenge the harsh self-talk with evidence from my life, and (5) offer specific, warm encouragement for moving forward. The tone should feel like a real friend, not a therapist: warm, honest, maybe even a little humor where appropriate.

πŸ’‘

Pro tip: Read this letter aloud to yourself. Hearing compassionate words in your own voice activates different neural pathways than reading silently.

Parts Work Dialogue

17/45

✨ What it does

ChatGPT guides a parts dialogue when part of you wants [desire 1] and another wants [desire 2], naming the tension in [describe the internal tension or paralysis]. You answer only if you want to, and you decide what you share.

I notice I have an inner conflict. Part of me wants to [desire 1] but another part of me wants to [desire 2]. This creates [describe the internal tension or paralysis]. Guide me through an Internal Family Systems-inspired exercise: (1) help me identify and name each part (e.g., The Protector, The Achiever, The Wounded Child), (2) interview each part: what is its role? what is it afraid will happen? what does it need?, (3) find the positive intention behind each part, even the ones that seem destructive, (4) facilitate a conversation between the parts, and (5) suggest how my core Self can honor all parts while making a decision.

πŸ’‘

Pro tip: Every part of you developed for a reason, usually to protect you from pain. Even destructive patterns started as survival strategies.

Radical Acceptance Practice

18/45

✨ What it does

ChatGPT walks a radical-acceptance exercise for a reality you cannot change, starting from the thought you keep repeating. You decide how much of the situation you describe, and you stop if it feels like pushing.

I am struggling to accept [describe a situation you cannot change: a loss, a limitation, a reality you keep fighting against]. I keep thinking: "[resistant thoughts, e.g., this should not have happened, it is not fair, I cannot live like this]." Walk me through a radical acceptance exercise: (1) name exactly what I am being asked to accept (be specific), (2) clarify what acceptance does and does not mean (it does not mean approval or giving up), (3) identify the cost of continuing to resist this reality, (4) write a series of acceptance statements I can practice ("It is true that..."), (5) suggest a physical ritual that symbolizes releasing resistance (e.g., open hands, exhale, placing a stone down), and (6) name what becomes possible AFTER acceptance that is impossible while fighting reality.

πŸ’‘

Pro tip: Acceptance is not a one-time event. You may need to practice it repeatedly for the same thing. That is normal, not failure.

Values Clarification Exercise

19/45

✨ What it does

ChatGPT walks a values exercise across life areas such as relationships, work, health, and community, then asks what a small aligned step would be. You rate only the areas you want to touch, and you pick one action or none.

I feel disconnected from what matters to me. I have been going through the motions and I want to realign my life with my values. Help me with a values clarification exercise: (1) present me with 10 life domains (relationships, career, health, creativity, spirituality, community, etc.) and for each, ask me to rate how important it is (1-10) and how aligned my current life is (1-10), (2) identify my top 5 values based on the gaps, (3) for each top value, help me define what it looks like in daily action (not abstract terms), (4) identify one small committed action I can take this week for each value, and (5) help me write a personal values statement I can revisit when I feel lost.

πŸ’‘

Pro tip: Values are directions, not destinations. You never arrive at "being kind": you practice it daily. Focus on direction, not achievement.

Shame Resilience Script

20/45

✨ What it does

ChatGPT walks shame-resilience questions about [describe what happened or what you are ashamed of] and the lines in [list shame thoughts]. You answer only what feels safe, and you decide what you share.

I am feeling intense shame about [describe what happened or what you are ashamed of]. The shame messages in my head are: "[list shame thoughts]." Help me work through this using BrenΓ© Brown's shame resilience framework: (1) recognize the physical sensations of shame (what I notice in my body), (2) identify the expectations or norms driving the shame ("I should..."), (3) reality-check these expectations: are they realistic? whose voice are they?, (4) connect me to shared humanity: who else has experienced something similar?, (5) write what I would say if a close friend told me they did/experienced the exact same thing, and (6) identify one person I could reach out to (shame thrives in secrecy).

πŸ’‘

Pro tip: Shame says "I am bad." Guilt says "I did something bad." If you can shift from shame to guilt, you can learn and grow instead of spiraling.

Relationship & Communication

5 prompts

Difficult Conversation Prep

21/45

✨ What it does

ChatGPT drafts an opening and a playbook for a hard talk with [person and relationship, e.g., partner, parent, boss, friend] about [topic], using [your key points]. You edit the words until they sound like you, then you decide whether to have the conversation.

I need to have a difficult conversation with [person and relationship, e.g., partner, parent, boss, friend] about [topic]. I have been avoiding it because [describe fear]. What I want to communicate: [your key points]. What I am afraid they will say or do: [describe]. Help me prepare: (1) write an opening statement using "I" language that is honest but not attacking, (2) anticipate their likely responses and prepare non-defensive replies, (3) identify my non-negotiables versus where I can be flexible, (4) create a script for if the conversation escalates (de-escalation phrases), and (5) plan what I will do after the conversation regardless of how it goes (self-care plan).

πŸ’‘

Pro tip: The goal of a difficult conversation is to be heard and to hear: not to win. Go in curious about their perspective, not just armed with your points.

Boundary Setting Script

22/45

✨ What it does

ChatGPT writes a clear, kind boundary statement for [person] about [situation], plus a line for when they push back. You change any sentence that does not sound like you, then you decide whether to say it.

I need to set a boundary with [person] regarding [situation]. This is hard for me because [describe why: guilt, fear of conflict, people-pleasing pattern, etc.]. The boundary I want to set: [describe what you want to change]. Write for me: (1) a clear, kind boundary statement (what I need, not what they are doing wrong), (2) an explanation of the consequence if the boundary is crossed (stated matter-of-factly, not as a threat), (3) responses to likely pushback (guilt-tripping, anger, dismissal, playing victim), (4) affirmations I can tell myself when guilt creeps in after setting the boundary, and (5) a reminder of why this boundary protects the relationship, not damages it.

πŸ’‘

Pro tip: A boundary is not a punishment for the other person. It is information about what you need to stay healthy in the relationship.

Attachment Style Reflection

23/45

✨ What it does

ChatGPT reflects on possible attachment patterns from how you described closeness, distance, and [describe] fears. You keep the framing that fits, and you ignore anything that does not sound like your relationships.

I want to understand my attachment style and how it affects my relationships. Here is how I typically behave in close relationships: - When my partner/friend pulls away, I tend to: [describe] - When someone gets too close, I tend to: [describe] - My biggest fear in relationships is: [describe] - I tend to attract people who are: [describe] - After conflict, I usually: [describe] Based on these patterns: (1) identify which attachment style(s) I likely lean toward and explain why, (2) trace these patterns to possible origins (without diagnosing), (3) show me how this style plays out in a specific relationship dynamic, (4) identify my attachment style's strengths (not just problems), and (5) suggest 3 practices to move toward more secure attachment.

πŸ’‘

Pro tip: Attachment style is not destiny. Earned secure attachment is possible at any age through awareness, healing relationships, and intentional practice.

Conflict Repair Template

24/45

✨ What it does

ChatGPT drafts a repair message after a conflict with [person] about [topic], owning [your part] without taking [their part]. You edit it until it is honest, then you decide whether to send it.

I had a conflict with [person] about [topic]. What happened: [describe the conflict]. I said/did: [your part]. They said/did: [their part]. I am feeling: [current emotions]. I want to repair this but I do not know how. Help me: (1) identify what I am responsible for (without taking blame for their part), (2) write a genuine apology for my part that does not include "but" or justifications, (3) express how I understand the impact of my actions on them (not just my intent), (4) suggest what I want to do differently next time (specific, not vague), and (5) make a request for what I need from them (not a demand). Write this as a message I could actually send or say.

πŸ’‘

Pro tip: A good apology focuses on their experience, not your reasons. "I understand that when I did X, you felt Y" lands better than "I did X because Z."

Codependency Pattern Check

25/45

✨ What it does

ChatGPT reviews the behaviors you list in your relationship with [person] and names which look like over-giving versus ordinary care. You decide what fits, and you pick one small change or none.

I suspect I might have codependent patterns in my relationship with [person]. Here are behaviors I have noticed in myself: [List 5-8 behaviors, e.g., putting their needs before mine, feeling responsible for their emotions, difficulty saying no, losing my identity, etc.] Help me understand: (1) which of these behaviors reflect codependency versus healthy caregiving, (2) what underlying belief drives each codependent pattern (e.g., "If I stop helping, they will leave"), (3) what healthy interdependence looks like in contrast to what I am doing, (4) one boundary or behavior change I could try this week for each pattern, and (5) signs that I am making progress toward healthier relating. Be direct with me, not coddling.

πŸ’‘

Pro tip: Codependency often masquerades as love or selflessness. The test: are you helping because you want to, or because you are afraid of what happens if you stop?

Go from copy-pasting to actually mastering AI.

AI Academy: hands-on tutorials on ChatGPT, Claude, Midjourney, and many other tools. New tutorials are added regularly.

Start Your Free Trial

Mindfulness & Journaling

5 prompts

Guided Journaling Session

26/45

✨ What it does

ChatGPT offers five journaling questions that start from [one-word emotion] and [brief description] and move a little deeper each time. You write only as far as you want, and you skip any prompt that feels like too much.

I want to journal but I do not know where to start. Today I am feeling: [one-word emotion]. The main thing on my mind is: [brief description]. Guide me through a 15-minute journaling session with 5 progressive prompts. Start surface-level (what happened today) and go deeper with each prompt until we reach something meaningful about my inner world. After each prompt, tell me to write for 3 minutes before moving to the next. Make the prompts specific to what I shared, not generic. End with a reflection prompt that ties everything together.

πŸ’‘

Pro tip: Do not edit or censor as you write. Stream of consciousness is the goal. You can always destroy the pages afterward.

Body Scan Meditation Script

27/45

✨ What it does

ChatGPT writes a 10-minute body-scan script that lingers on [areas, e.g., jaw, shoulders, stomach] at stress level [1-10]. You read or record it, then follow only the parts that feel safe.

Write a 10-minute body scan meditation script I can read to myself or record. I tend to hold tension in [areas, e.g., jaw, shoulders, stomach]. My current stress level: [1-10]. Customize the script to: (1) start with 5 deep breaths with specific counts, (2) move systematically from feet to crown, (3) spend extra time on my tension areas with specific release techniques, (4) include curiosity-based language ("notice without judging") rather than commands, (5) incorporate a brief compassion element (sending kindness to areas of pain), and (6) end with full-body awareness and a gentle return to the room. Use calming, unhurried language throughout.

πŸ’‘

Pro tip: Record yourself reading this slowly and use it as a guided meditation. Your own voice is surprisingly effective for relaxation.

Morning Intention Setting

28/45

✨ What it does

ChatGPT builds a short morning ritual from [describe] mornings, [describe] struggle, and a focus such as [general intention, e.g., patience, courage, presence, boundaries]. You try the breathing and questions, and you drop what does not fit.

Help me create a morning intention-setting ritual. My mornings currently look like: [describe]. My biggest struggle right now: [describe]. Today I want to focus on: [general intention, e.g., patience, courage, presence, boundaries]. Create: (1) a 2-minute breathing exercise to start, (2) 3 reflection questions to ask myself each morning, (3) a customized intention statement for today based on what I shared, (4) one micro-action (under 5 minutes) that embodies this intention, and (5) an evening check-in question to reflect on how the intention went. Make it realistic for someone who is not a morning person.

πŸ’‘

Pro tip: Keep it under 10 minutes total. A short ritual you actually do is infinitely better than an elaborate one you abandon after 3 days.

Therapeutic Letter to Younger Self

29/45

✨ What it does

ChatGPT drafts a letter from you now to your [age]-year-old self, about [describe what was happening] and [your sense of what was missing]. You change any line that does not sound like you, and you decide whether anyone else ever sees it.

I want to write a letter to my [age]-year-old self. At that age, I was going through: [describe what was happening]. What I needed to hear but nobody told me: [your sense of what was missing]. Write a letter from my current self to that younger version of me. The letter should: (1) describe what I see when I look at that younger self (with tenderness), (2) acknowledge specifically what was hard and validate that it was not okay, (3) tell them what I wish someone had said, (4) share what I know now that would have helped then, (5) let them know how things turn out (without minimizing the pain of getting here), and (6) thank them for surviving and getting us to where we are now.

πŸ’‘

Pro tip: If this brings up tears, let them come. Grief for your younger self is a sign of compassion, not weakness.

Weekly Therapy Reflection

30/45

✨ What it does

ChatGPT helps you put [brief summary of session topics] and [key insight or observation] into your own words, then suggests a question for next week. You keep the takeaway that is yours, and you decide what you bring back to therapy.

I want to process my therapy session from this week. What we discussed: [brief summary of session topics]. What my therapist pointed out: [key insight or observation]. How I felt during the session: [emotions]. Something that surprised me: [if anything]. Help me: (1) identify the key takeaway and put it in my own words, (2) notice any resistance or defensiveness I felt and what it might mean, (3) connect what was discussed to patterns in my daily life this week, (4) suggest one thing I can practice before my next session, and (5) write down what I want to bring up next session based on what came up today.

πŸ’‘

Pro tip: The real work of therapy happens between sessions. Use this after every session to double the impact of your investment.

Trauma-Informed Exercises

5 prompts

Window of Tolerance Assessment

31/45

✨ What it does

ChatGPT explains a window of tolerance in plain language, then helps you map calm, speed-up, and shut-down from [describe] and the body signs you name. You use the map only if it fits, and you decide what you share.

I want to understand my window of tolerance. Describe what it means, then help me map mine. When I am in my window (regulated), I feel: [describe]. Signs I am moving into hyperarousal (fight/flight): [describe symptoms: racing heart, irritability, panic, etc.]. Signs I am moving into hypoarousal (freeze/shutdown): [describe symptoms: numbness, dissociation, exhaustion, etc.]. My common triggers: [list]. Based on this, create: (1) a personal window of tolerance map with my specific signs, (2) 3 strategies for when I notice hyperarousal, (3) 3 strategies for when I notice hypoarousal, (4) a daily practice to widen my window over time, and (5) a simple scale I can use to check in with myself throughout the day.

πŸ’‘

Pro tip: The goal is not to never leave your window. The goal is to notice sooner and return faster. This gets easier with practice.

Safety and Resource Building

32/45

✨ What it does

ChatGPT drafts a safe-place picture and a container image you can use when feelings get big, before any deep trauma work. You change the details until the place feels yours, and you stop if it does not help.

I want to build my internal resources for managing difficult emotions and trauma responses. Help me create: (1) a detailed safe place visualization: guide me through creating an imaginary place where I feel completely safe (what I see, hear, feel, smell), (2) a container exercise: help me visualize a container where I can place distressing thoughts and images when I need to function (what it looks like, how I seal it, where I put it), (3) a resource inventory: list my existing coping resources across categories (people, places, activities, internal strengths, sensory tools), and (4) a personalized self-soothing kit for each sense (touch, smell, taste, sight, sound).

πŸ’‘

Pro tip: Practice the safe place visualization when you are calm so it is accessible when you need it. It takes 15-20 repetitions to become automatic.

Trigger Response Plan

33/45

✨ What it does

ChatGPT drafts an early-warning and aftercare plan for a trigger you name, including what to do in the body before the full reaction. You keep the steps that help, and you decide how much of the trigger you describe.

I keep getting triggered by [describe trigger: a situation, sensory experience, person, topic, etc.]. When triggered, my reaction is: [describe what happens in body, emotions, thoughts, and behavior]. Help me build a trigger response plan: (1) early warning signs that I am being triggered (before the full response), (2) immediate grounding steps I can take in the first 30 seconds, (3) a script I can say to myself that orients me to the present ("I am safe. I am in [place]. It is [year]. That was then, this is now."), (4) what to do if I am with other people when triggered (exit strategy, brief explanation I can give), and (5) a self-care protocol for after the trigger response passes.

πŸ’‘

Pro tip: Share this plan with a trusted person so they can help you use it when you are too activated to remember the steps.

Narrative Coherence Exercise

34/45

✨ What it does

ChatGPT helps you line up fragments of a hard experience in time, without forcing a full story. You share only what feels safe, and you stop whenever you want.

I want to work on creating a coherent narrative about a difficult experience. The experience: [describe briefly: you can be as vague or specific as you are comfortable with]. Current fragmented pieces: [what you remember: facts, emotions, body sensations, beliefs formed]. Help me: (1) organize these fragments chronologically (what happened before, during, after), (2) identify gaps where memory is unclear and normalize this, (3) add the emotional layer: for each phase, what was I likely feeling and why does that make sense, (4) identify the meaning I made from this experience (beliefs about self, others, the world), and (5) explore whether that meaning still serves me or needs updating. IMPORTANT: keep this at a manageable level: we are organizing, not reliving.

πŸ’‘

Pro tip: Stop immediately if you feel overwhelmed. This exercise should feel like organizing a closet, not reliving a storm. Titrate the intensity.

Post-Traumatic Growth Exploration

35/45

✨ What it does

ChatGPT asks honest growth questions about [describe difficult experience or period], without silver linings or a claim that it was meant to be. You answer only what feels true, and you skip any question that feels forced.

I have been through [describe difficult experience or period]. I am past the acute phase and beginning to heal. I want to explore whether any growth has come from this experience: without toxic positivity or pretending it was "meant to be." Help me reflect honestly: (1) what did this experience teach me about myself that I could not have learned any other way?, (2) how have my relationships or priorities shifted since?, (3) have any new possibilities or paths opened that were not visible before?, (4) has my sense of personal strength changed?, (5) has my philosophy of life or spirituality evolved? For each area, help me distinguish between genuine growth and performative "everything happens for a reason" narratives. It is okay if some areas show no growth: that is honest.

πŸ’‘

Pro tip: Post-traumatic growth and ongoing pain can coexist. You do not have to be grateful for what happened to acknowledge that you grew.

Depression Support

5 prompts

Low Energy Day Survival Plan

36/45

✨ What it does

ChatGPT drafts a low-energy survival plan for today at energy [1-10], triaging [list any obligations] into what cannot wait and what can. You do the smallest version, and you ignore any step that sounds like a productivity list.

I am having a very low energy day. My depression is making everything feel impossible. My current energy level: [1-10]. Things I have to do today that I cannot skip: [list any obligations]. Help me create a survival plan for today: not a productivity plan, a survival plan. (1) Triage my obligations: what truly cannot wait versus what I am telling myself cannot wait? (2) Create the absolute minimum version of each necessary task (e.g., "shower" becomes "wash face with a wet cloth"). (3) Build in 3 micro-comforts throughout the day (tiny things that take no energy but provide a moment of relief). (4) Write me a permission slip to not do the things that can wait. (5) Set one tiny goal for the day that I can feel good about accomplishing, even if it seems pathetically small to a non-depressed person.

πŸ’‘

Pro tip: On depression days, brushing your teeth is an achievement. Redefine success for today. Tomorrow can have different standards.

Motivation When Nothing Matters

37/45

✨ What it does

ChatGPT names possible reasons nothing feels meaningful, then offers tiny next steps around [things you used to care about] that do not require feeling motivated. You try one small action, or none, and you decide what you share.

I am in a phase where nothing feels meaningful or motivating. I know I "should" care about [things you used to care about] but I feel flat and empty. I am not in crisis: just stuck in the gray. Help me: (1) normalize this experience (is this anhedonia? burnout? depression?), (2) instead of motivation, help me identify 3 tiny actions I can take based on routine or commitment rather than feeling (acting opposite to the mood), (3) identify one sensory experience I could try today to wake up my system even slightly (cold water, strong flavor, loud music, physical movement), (4) write me a message from my future self who got through this gray period, and (5) create a "spark list": 20 tiny activities that have even a 5% chance of producing a flicker of interest or pleasure.

πŸ’‘

Pro tip: You do not need to feel motivated to act. Act first, and sometimes feeling follows. This is the core principle of behavioral activation.

Automatic Negative Thought Log

38/45

✨ What it does

ChatGPT examines [thought 1] through [thought 5], names the thinking trap, and offers a fairer rewrite plus how much the old thought still feels true. You keep the rewrites that fit, and you leave the ones that do not.

I have been having a lot of negative thoughts lately. Here are the main ones running on repeat: 1. "[thought 1]" 2. "[thought 2]" 3. "[thought 3]" 4. "[thought 4]" 5. "[thought 5]" For each thought: (1) name the cognitive distortion it represents, (2) rate how much I believe it right now (0-100%), (3) provide evidence against it from my own life, (4) write a balanced alternative thought that is realistic (not just positive), (5) rate how much I believe the alternative (0-100%). Then identify which thought is the most damaging to my daily functioning and suggest one behavioral experiment I could do this week to test whether that thought is actually true.

πŸ’‘

Pro tip: You do not need to believe the alternative thought immediately. Repetition over weeks gradually shifts belief. Think of it as retraining, not convincing.

Pleasurable Activity Rediscovery

39/45

✨ What it does

ChatGPT suggests three easier versions of [list activities you used to like] when they now feel flat or pointless. You try one tiny version, then notice how you felt, without forcing enjoyment.

Depression has stolen my ability to enjoy things. I used to enjoy: [list activities you used to like]. Now they all feel: [describe: pointless, exhausting, boring, etc.]. Help me rebuild my relationship with pleasure: (1) from my old list, identify 3 activities to try again but in a modified, low-barrier version (e.g., "hiking" becomes "walk to the end of the block"), (2) suggest 5 completely new micro-activities I have never tried that require minimal energy, (3) for each activity, set the bar absurdly low (do it for 2 minutes, not an hour), (4) create a simple tracking sheet where I rate mood before and after each activity (even if both ratings are low), and (5) remind me that I do not have to enjoy these activities: I just have to do them and observe. Enjoyment often returns after the behavior, not before.

πŸ’‘

Pro tip: Rate your mood 1-10 before and after each activity. Over time, you will have data showing which activities move the needle, even slightly.

Depression Communication Script

40/45

✨ What it does

ChatGPT drafts an honest script for telling someone about your depression, including what they already know and what you need them to understand. You edit it until it sounds like you, then you decide whether to say it.

I need to tell someone in my life about my depression. The person: [who: partner, parent, friend, boss, etc.]. What they currently know: [what they have noticed or been told]. What I want them to understand: [what you need them to know]. What I am afraid of: [their reaction]. Help me: (1) write an opening that is honest but not alarming, (2) describe my experience in concrete terms they can understand (not clinical language), (3) tell them specifically what helps and what does not help (e.g., "Please do not tell me to exercise" or "It helps when you just sit with me"), (4) set expectations about what this will look like day to day, and (5) let them know it is not their job to fix it. Write this as a letter or script I can read to them if speaking is too hard.

πŸ’‘

Pro tip: Telling someone is an act of trust, not burden. Most people want to help but do not know how: giving them specific guidance is a gift.

Stress & Burnout

5 prompts

Burnout Assessment and Recovery Plan

41/45

✨ What it does

ChatGPT helps sort burnout from ordinary stress or low mood using [describe workload, boundaries, support, etc.] and the symptoms you name, then sketches a recovery plan. You keep the boundary that is realistic, and you treat this as a draft, not a diagnosis.

I think I might be burned out. Here are my symptoms: [describe: exhaustion, cynicism, reduced efficacy, physical symptoms, etc.]. My work/life situation: [describe workload, boundaries, support, etc.]. Help me: (1) assess whether this is burnout, stress, or depression (explain the differences), (2) identify the specific dimensions of burnout I am experiencing (emotional exhaustion, depersonalization, reduced accomplishment), (3) pinpoint the top 3 contributing factors I might be able to change, (4) create a 30-day recovery plan with weekly milestones that is realistic given my constraints, and (5) identify the minimum viable boundaries I need to set immediately to stop the bleeding.

πŸ’‘

Pro tip: Burnout is a response to your environment, not a personal failing. Recovery requires changing conditions, not just coping harder.

Stress Inventory and Triage

42/45

✨ What it does

ChatGPT sorts a full stress list into what you can control, what you can partly control, and what you cannot, then suggests a first action. You add the real items, then pick one controllable step or none.

I am overwhelmed by everything on my plate. Here is my complete stress inventory: [List everything causing you stress: work, relationships, health, finances, family, etc.] Help me triage these: (1) categorize each stressor as: within my control, partially in my control, or completely outside my control, (2) for items in my control, identify one action step for each, (3) for items partially in my control, identify what I can influence and what I need to release, (4) for items outside my control, write an acceptance statement for each, (5) identify the top 3 stressors that would create the most relief if addressed, and (6) create a "not right now" list: things I formally give myself permission to table for the next 30 days.

πŸ’‘

Pro tip: The "not right now" list is the most powerful part. Actively deciding to defer something reduces its mental load even though nothing externally changed.

Compassionate Boundary Audit

43/45

✨ What it does

ChatGPT estimates time for [number] real hours in a week against every commitment you list, then shows where you are overbooked. You correct the time guesses, then choose one thing to drop or shrink.

I keep overcommitting and then burning out. My current commitments: [list everything: work, social, family, volunteer, etc.]. My available hours per week (realistically, including rest): [number]. Help me: (1) estimate the actual time each commitment requires (I probably underestimate), (2) show me the math: am I overcommitted? by how much?, (3) categorize commitments as essential, important, nice-to-have, and should-not-be-mine, (4) identify 3 commitments I could reduce or drop with the least consequences, (5) write a kind but firm script for each one to scale back, and (6) establish a rule for how I will evaluate future requests before saying yes.

πŸ’‘

Pro tip: Every yes is a no to something else. Usually, that something else is your health, rest, or the people closest to you.

Nervous System Reset Routine

44/45

✨ What it does

ChatGPT writes a 20-minute daily reset for a body stuck in overdrive since [timeframe], with breathing and simple downshift steps. You try the pieces that help, and you stop if anything feels worse.

My nervous system has been stuck in overdrive for [timeframe]. I am constantly in fight-or-flight mode. Symptoms: [describe: muscle tension, poor sleep, irritability, digestive issues, racing thoughts, etc.]. Create a daily nervous system reset routine I can do in 20 minutes: (1) a vagus nerve stimulation exercise (with explanation of why it works), (2) a specific breathing technique for activating the parasympathetic system, (3) a gentle movement or stretch sequence for releasing stored tension, (4) a bilateral stimulation exercise (butterfly taps or eye movements), and (5) a closing ritual that signals safety to my body. Include timing for each step and brief physiological explanations so I understand why each one works.

πŸ’‘

Pro tip: Your body keeps the score. Chronic stress is stored physically. Talking about stress helps, but physical practices are needed to release it from the body.

Emergency Self-Care Protocol

45/45

✨ What it does

ChatGPT writes an immediate self-care plan for the next [amount of time], starting with a 60-second ground, for a hard moment that is not an emergency. You use only the steps that help, and you get human help if you are not safe.

I am in a tough spot emotionally right now. I am not in danger, but I am not okay. I have [amount of time] before I need to function again. My current state: [describe: overwhelmed, crying, panicking, numb, angry, etc.]. Create an emergency self-care protocol for right now: (1) a 60-second grounding exercise I can do immediately, (2) what my body probably needs right now (water, food, temperature change, movement, stillness), (3) one physical action to do with my hands (tactile grounding), (4) a brief mantra or phrase customized to what I am feeling, (5) permission to do the bare minimum for the rest of the day, and (6) a plan for the next 2 hours that prioritizes safety and gentleness. No toxic positivity. Just practical, warm guidance for getting through the next few hours.

πŸ’‘

Pro tip: Bookmark this response. When you are in this state again, you will not have the energy to write a prompt. Having the protocol ready is the point.

Free tool

Prompt Optimizer

Turn a rough idea into a structured, professional AI prompt.

Try it free β†’

Frequently Asked Questions

No. ChatGPT cannot diagnose conditions, provide clinical treatment, or respond to crises. It lacks the attunement, ethical training, and relational depth of a human therapist. These prompts are designed as supplements to professional care: tools for practicing therapeutic skills between sessions, processing emotions through writing, and building self-awareness. If you are in crisis, contact the 988 Suicide and Crisis Lifeline (call or text 988) or your local emergency services.
Yes. The prompts draw from established evidence-based approaches including Cognitive Behavioral Therapy (CBT), Dialectical Behavior Therapy (DBT), Acceptance and Commitment Therapy (ACT), Internal Family Systems (IFS), and self-compassion research by Kristin Neff. However, using a technique in ChatGPT is not equivalent to receiving therapy. The techniques are most effective when guided by a trained professional who can adapt them to your specific situation.
Be mindful about what you share. ChatGPT conversations may be reviewed by OpenAI for training purposes unless you opt out in settings. Avoid sharing identifying details about other people, specific locations, or information that could be harmful if exposed. You can use vague descriptions instead of specific details and still get useful results. For highly sensitive material, consider using the prompts as templates and writing your full responses in a private journal instead.
Quality matters more than frequency. Using 2-3 prompts per week with genuine reflection is more valuable than rushing through all of them. Many therapists recommend journaling 3-4 times per week, and these prompts work well on that schedule. If you are in active therapy, consider using relevant prompts after each session to deepen your processing. If any prompt consistently increases your distress rather than helping you process it, skip it and discuss the topic with a professional instead.
Stop the exercise immediately. Use the grounding techniques from the Anxiety Management section to return to the present. Remember that you are safe: processing difficult emotions in writing can feel intense, but you are in control. If the distress does not subside, reach out to a trusted person or contact a crisis line. This is a sign that the material needs professional support, not that something is wrong with you. Note which prompt triggered the response and bring it to your next therapy session: it is valuable clinical information.

Prompts are the starting line. Tutorials are the finish.

A growing library of hands-on tutorials on ChatGPT, Claude, Midjourney, and many other AI tools. New tutorials are added regularly.

7-day free trial. Cancel anytime.