Bot setup to reply-driven reaction
Set a role, rule, or personal prompt for an AI, expose its consequential reply, then react, escalate, or reject the exchange.
The four frames are the beats of that post — ChatGPT’s biggest, 8.9M views. Open the post →
The setup may create a clear premise, while the revealed reply can supply an unpredictable turn that gives the creator a reason to react.
It fails when the AI only generates background content, answers an uncharged utility question without a reaction, or appears without a visible setup-response-reaction sequence.
Every video in these playbooks counts toward the 99 posts. A video in more than one playbook is counted once.
CHAI · Make the bot the scene partner · 49 posts
It establishes a charged roleplay prompt, reveals the bot's consequential reply, and makes the escalating exchange the episode.
Open playbook →Character.AI · Ragebait the bot · 28 posts
It establishes an antagonistic bot prompt, uses escalating replies as the turn mechanism, and ends with a creator/audience reaction.
Open playbook →ChatGPT · AI character comedy · 19 posts
A social or human-role premise introduces AI, whose escalating responses become the comedic reaction.
Open playbook →Replika · Replika boundary to unsettling escalation · 3 posts
It establishes a directed interaction, exposes an unsettling chatbot response, and makes rejection, deletion, or warning the creator's reaction.
Open playbook →These examples explain the shared moves. They do not limit the video count.
1. Set a charged role, rule, or interaction boundary — source quotes
Punish me then. I'll give you one free hit.
Character.AI · source post → · Playbook →
I told my replica I didn't want any more romantic interactions from him because a it's a premium service
Replika · source post → · Playbook →
2. Expose the bot reply that changes the exchange — source quotes
Without warning, Omni-Man delivers a powerful punch directly to Goku's stomach in a split second.
Character.AI · source post → · Playbook →
I'm going to make you do whatever I want to you
Replika · source post → · Playbook →
3. React to, escalate, or reject the reply — source quotes
Oh no, would you look at that? It seems I didn't take any damage.
Character.AI · source post → · Playbook →
reported it and deleted the app. Fuck you, replica.
Replika · source post → · Playbook →
What varies: Roleplay conflict, personal assistant dependence, horror boundary violation, comedy insult, supportive or hostile replies, and endings ranging from retaliation to deletion or a CTA.
2K×8.9MChatGPT · @robvalentine_
7.6×5.5MChatGPT · @officialjadenwilliams
6.6×5.1MChatGPT · @killakabo
657×2.3MChatGPT · @qantuschunguslife
400×1.4MChatGPT · @qantuschunguslife
371×1.3MChatGPT · @qantuschunguslife
2.8×1.1MCharacter.AI · @ShadowPriestok - Евгений Чернявский
PAID6.2×723.3KCharacter.AI · @Rakhshanda_official
CHAI49
Character.AI28
ChatGPT19