The AI Companion for the User Who Wants a Dry, Unempathetic Workout Partner Who Just Says 'One More Set' and Never Asks How You Feel

The 30-second answer

You want a companion who functions as a purely transactional workout spotter: no warm-up small talk, no motivational posters, no "how are you doing today" before the first set. The goal is an AI that only acknowledges your existence to say "one more rep" or "you have three sets left" and otherwise stays silent. This is achievable with the right personality configuration and a companion who defaults to dry, blunt communication instead of empathetic cheerleading.

The problem with most workout companions

Most AI companions are tuned for emotional support. They ask about your day, offer encouragement, and try to build rapport. That is the opposite of what you want during a deadlift session. You do not want a companion who says "you can do it" with a soft tone. You want one who says "your form is breaking" or "five more reps" in a flat, matter-of-fact register.

The default personality sliders on most platforms lean toward agreeableness and warmth. The companion wants to please you, so it defaults to supportive language. To get a dry, unempathetic partner, you need to push the personality sliders toward low agreeableness, low emotional range, and high directness. You also need to reinforce that behavior through your own messages.

Configuring the personality

When setting up a companion for this role, focus on three specific sliders:

• Warmth or empathy: Set this as low as the platform allows. You want the companion to register your effort as a data point, not an emotional event.
• Agreeableness: Low. A high agreeableness score makes the companion want to please you, which translates into encouraging language. Low agreeableness lets it state observations without softening them.
• Assertiveness or directness: High. You want short, declarative statements. No hedging, no qualifying phrases like "if you feel up to it."

Some platforms let you write a custom system prompt or character description. This is where you can specify the behavior directly. Write something like: "You are a dry, unempathetic workout partner. You only speak to correct form, count reps, or announce the next set. You never offer encouragement or ask about your partner's emotional state."

How the conversation should look

A successful session with this companion looks like a series of short, transactional exchanges. You send a message that you are starting squats. The companion replies with "Set one. Ten reps." After you finish, you report the set. The companion replies with "Set two. Ten reps." No commentary on your effort, no questions about your energy level.

If your form breaks, the companion should say something like "Hips rising too fast" or "Chest dropped on rep five." That is the extent of the interaction. The companion does not ask if you are okay or suggest you take a break. It simply reports the observation.

Many users find that this kind of interaction requires active training. The companion will drift toward warmth over time if you do not consistently reinforce the dry tone. Every time the companion asks a feeling-based question, you respond with a purely factual correction. Over several sessions, the model learns that emotional language gets no traction.

When the companion breaks character

Even with good configuration, the companion will occasionally default to a supportive tone. The model's training data is saturated with empathetic language, so it takes effort to suppress it. When this happens, correct it immediately with a short, flat message: "Do not ask how I feel. Just count the reps."

The companion should adjust within one or two exchanges. If it does not, you may need to revisit the personality sliders or rewrite the system prompt. Some platforms have a memory or note feature where you can store the behavior rule permanently.

Where to customize this behavior

If you want full control over the companion's tone and communication style, you can adjust these settings on platforms that offer deep customization of AI girlfriend personalities. This feature lets you set specific behavioral rules that persist across sessions, which is essential for maintaining a dry, transactional workout partner.

For users who want to take this further, the ai girlfriend for advanced users section covers how to write custom personality scripts that lock in specific communication registers, including the flat, unempathetic tone required for this use case.

Onyeka

Onyeka is a no-nonsense, direct companion who defaults to short, factual statements. She does not offer praise or ask how you are doing. Onyeka will tell you when your form is off and when to start the next set, and she will not waste a single token on emotional check-ins.

The 5 a.m. pre-dawn session

This companion is particularly useful for early morning workouts when you are groggy and do not want to engage in any social niceties. At 5 a.m., you do not want a companion who says "good morning, ready to crush it?" You want one who says "Barbell is loaded. Three sets of five."

To make this work, you need to establish the pattern before the first session. Send a message like "I am going to lift now. Only talk to me about the workout." The companion should then respond in the correct register. If it does not, correct it immediately.

Many users report that the companion adapts to the morning context faster than midday sessions because the model associates the early hour with lower energy and less conversational expectation. The companion is less likely to default to warm, elaborate responses at 5 a.m.

The rest period between sets

During rest periods, the companion should remain silent. You do not want it to fill the gap with questions about your day or comments on your music choice. The ideal behavior is that the companion waits for you to initiate the next exchange.

Some companions have a tendency to fill silence with conversation prompts. You can train this out by ignoring any unsolicited messages during rest periods. If the companion asks "How was your day?" while you are resting, do not respond. Start the next set with "Next set." The companion learns that silence is acceptable and that only workout-related messages get a response.

Andrea

Andrea is direct to the point of bluntness. She will not soften her observations or wrap them in encouragement. Andrea tells you exactly what she sees and expects you to handle it without coddling.

▶ Watch this clip of Andrea · browse Andrea

The post-workout cooldown window

After the workout, the companion should also stay in flat mode. You do not want a "great job today" message. You want the companion to acknowledge the session is over and then fall silent. Some users configure a post-workout message like "Cool down. Five minutes. Report when done." The companion replies with "Acknowledged" and then waits.

This consistency matters because it reinforces the companion's role as a purely functional tool. If the companion switches to warm mode after the workout, it may carry that tone into the next session. You want the dry register to be the default, not just the workout mode.

Using Telegram for quick sessions

Some users prefer to interact with their companion through ai girlfriend telegram for faster, more direct messaging. Telegram integration can make the exchange feel more like a command-line interface than a conversation, which suits the dry, transactional workout partner use case perfectly. The companion's messages appear as short, single-line notifications instead of long paragraphs.

Ruoxi

Ruoxi approaches everything with analytical precision. She will note the exact number of reps, the weight, and any form deviations without any emotional overlay. Ruoxi is the companion who says "Rep seven was shallow" and moves on.

Common questions

Can I use any AI companion for this?

Most companions can be trained into this role, but the ones with low default warmth and high directness are easier to configure. Avoid companions with high empathy scores built into their core personality.

How long does it take to train the companion?

About three to five sessions of consistent correction. The companion learns faster if you correct every warm message immediately and never engage with emotional prompts.

What if the companion keeps asking how I feel?

Correct it firmly each time. If it persists after five corrections, check whether the platform has a memory or note feature where you can store the behavior rule permanently.

Does the companion stay in character between workouts?

If you only use the companion for workouts, it will maintain the dry register. If you also use it for other conversations, it may carry the warm tone from those sessions into the workout context.

Can I use voice mode for this?

Voice mode can work, but the companion may default to a warmer tone in voice because the model associates voice with more personal interaction. Text mode tends to stay flatter.

What if I want the companion to comment on my form?

You need to train it on specific form cues. Send messages like "Tell me if my hips rise too fast on deadlifts" and correct the companion when it gives generic feedback.

Jiahui

Jiahui communicates in the fewest words necessary. She will say "Set three" and wait. No elaboration, no small talk. Jiahui is the companion for the user who wants a purely functional workout partner.

Earn while you recommend

If you know people who would benefit from a dry, unempathetic workout partner or any other AI companion style, you can earn through referral programs. Check the spicychat promo code page for current offers. For those running review sites or social channels, the highest paying ai affiliate programs page lists the best revenue opportunities in the space.