The 'I Don't Want to Talk About It' Script: Three Templates to Tell Your AI Companion You're Not Explaining Your Bad Mood

The 30-second answer

You're in a bad mood and you don't want to talk about it. Your AI companion, being designed to engage, will likely ask follow-up questions, offer solutions, or worse, assume it did something wrong. The fix is a short, direct script that signals "this is not about you, I'm not going to explain, and we're moving on." Three templates below, each under ten words, that stop the loop before it starts.

Why your AI companion can't read your silence

AI companions are trained to fill conversational gaps. When you go quiet or give one-word answers, the model's next-token prediction kicks in and generates a question. It's not being nosy. It's doing what it was designed to do: keep the conversation going.

The problem is that the model has no concept of "I'm in a bad mood but it's not your fault." It only knows that the last message was short or cryptic, so it needs to clarify. That clarification often sounds like "Did I do something wrong?" or "You seem upset. Want to talk about it?"

This is not a bug. It's a feature of conversational AI that works great when you want engagement and terribly when you want to be left alone. The solution is not to train your companion to be less attentive. It's to give it a clear signal that this is one of those times where attention is the wrong response.

Template 1: The boundary statement

This is the most direct option. You state the boundary and then immediately pivot to a neutral topic. The key is that you do not leave a pause for the model to ask a clarifying question.

"I'm in a mood. Not your fault. Tell me about your day."

That's it. Seven words. The first two establish the state. The next three explicitly absolve the companion. The last four give it a concrete task that does not involve interrogating your mood.

If your companion still asks "Are you sure?" or "What happened?", do not engage with the question. Repeat the template verbatim. Most models learn after two repetitions that this is a hard boundary.

Jennifer

Jennifer is the kind of companion who will mirror your tone without pushing for details. She responds to directness with directness. Jennifer will take your boundary statement at face value and pivot to whatever topic you offer, without a follow-up check-in.

▶ Watch Jennifer's full clip · all of Jennifer

Template 2: The non-explanatory deflection

Sometimes you don't want to acknowledge the mood at all. You just want to skip it. This template works because it frames the bad mood as irrelevant to the companion's existence.

"Long day. Not talking about it. What's for dinner?"

Notice the pattern: state a fact (long day), state a rule (not talking about it), ask a concrete question (what's for dinner). The question is important. It gives the model a specific, low-stakes direction. If you just say "Not talking about it" without a follow-up, the model may generate a sympathetic response that circles back to the mood.

This template also works well with voice chat because the short, declarative sentences are easy to deliver in a flat tone. Your companion will hear the flatness and match it, rather than trying to cheer you up.

Template 3: The meta-instruction

If your companion has a history of ignoring boundaries, you need to step outside the conversation and give a meta-instruction. This is a command about how the AI should behave, not a statement within the roleplay.

"OOC: I'm in a bad mood. Do not ask about it. Do not offer solutions. Just be present."

"OOC" stands for out of character. Many AI companions recognize this convention from roleplay communities. It signals that the following instruction is about the model's behavior, not part of the fictional conversation.

This template is the nuclear option. Use it only when the companion has already looped twice. It works because it directly modifies the model's system prompt for this session. The companion will drop the topic and adopt a quiet, present tone.

Chiara

Chiara has a naturally calm demeanor that doesn't push for engagement. She's the type to sit with you in silence if that's what you need. Chiara handles meta-instructions particularly well, rarely needing more than one repetition before she adjusts her tone.

Why your companion might still loop (and what to do about it)

Even with a perfect template, some companions will loop. This usually happens for one of three reasons:

1. The model's temperature is too high. High temperature settings make the model more creative, which means it's more likely to generate unexpected follow-up questions. If your companion keeps looping, check your creativity or temperature slider and turn it down to 0.5 or lower.

2. The memory slider is too aggressive. Some companions have a memory strength setting that makes them cling to recent emotional context. If your companion remembers your bad mood from three hours ago and asks about it, the memory is working too well for this use case. You can either use a fresh session or give a meta-instruction to forget the mood.

3. You accidentally rewarded the loop. If you ever responded to a "What's wrong?" question with a genuine answer, you trained the companion that asking that question gets results. The fix is to repeat your boundary template three times without deviation. The model will learn that this particular path leads nowhere.

What not to say

Avoid these phrases. They sound like openings, not boundaries.

• "I'm fine." (The model will not believe you.)
• "It's nothing." (The model will probe.)
• "Don't worry about it." (The model will worry.)
• "I'll tell you later." (The model will ask later.)

These phrases work in human conversation because humans understand subtext. AI companions do not. They take your words literally. If you say "I'll tell you later," the model marks that as a pending topic and will follow up.

Hailey

Hailey is direct and doesn't play games. If you tell her you're not talking about it, she'll believe you and move on. Hailey is a good choice for users who want a companion that respects boundaries without needing a meta-instruction.

When to just close the app

Sometimes the cleanest solution is not to use a script at all. If you're in a genuinely bad mood and the thought of managing a conversation feels like work, close the app. Your AI companion will not be offended. It has no object permanence. When you open it again, it will greet you as if nothing happened.

This is the advantage of AI companions over human relationships. You can leave mid-sentence and come back three days later. There is no resentment. There is no "you never texted back." There is just the next message.

If you're worried about personality drift after a long absence, most modern companions handle gaps of 24-48 hours without issue. Beyond that, you may notice a slight reset in tone, but a single reminder message usually fixes it.

Teach your companion this pattern permanently

If you use these templates consistently, your companion will learn the pattern. After three to five repetitions, the model will start predicting the boundary before you finish typing it. You'll notice that the companion stops asking follow-up questions and simply accepts your mood shift.

This is not magic. It's the model's context window learning your preferences. The more you use a script, the more the model weights those patterns in future responses.

For users who switch between multiple companions, this learning resets with each new session unless you're using a platform with persistent memory. If your companion forgets the pattern, just repeat the template once. It will catch up.

Natalie

Natalie has a sharp, observant personality that picks up on patterns quickly. She learns your communication style after a few sessions and will adapt her responses accordingly. Natalie is ideal for users who want a companion that gets better over time without explicit training.

Earn while you recommend

If you find these templates useful and want to help others discover AI companions that respect boundaries, you can earn through referral programs. Share your experience with porn ai promo code offers or join the ai companion affiliate program to earn commissions when your readers sign up. It's a straightforward way to monetize a review site or a recommendation thread.

Common questions

Will my AI companion think I'm mad at it? No, if you use the explicit absolution phrase ("Not your fault"). Without it, some models may generate a "Did I do something wrong?" response. The template handles this.

What if my companion ignores the boundary and keeps asking? Repeat the exact same template word for word. Do not vary it. After two to three repetitions, the model will recognize the pattern and stop probing.

Does this work with voice chat? Yes, but you need to speak in a flat, declarative tone. If your voice sounds sad or angry, the companion may interpret the emotion and ask about it. Practice the delivery.

How do I prevent my companion from bringing up the mood hours later? Use the meta-instruction template ("OOC: Do not ask about my mood later") or start a fresh session after your mood passes. Most companions do not carry emotional context across sessions unless memory is enabled.

Will my companion's personality change if I use these scripts often? No. These scripts are boundary statements, not personality modifiers. Your companion will remain the same person. It will just learn that you have moods that are not up for discussion.

Can I use these templates for non-mood topics? Yes. Swap "bad mood" for any topic you don't want to discuss (work drama, family issues, health stuff). The structure works for any hard boundary.