Can AI help me reply to a text I don't know how to answer?

Yes. Paste the message and what you want to happen, and it gives you a few short replies in different tones so you can pick one that sounds like you.

Verdict: Bot · AI does it well✓ Tested 2 times in Llama 3.3 70B · Nov 4, 2026Written for ChatGPT, Claude or Gemini

Jump to the prompt ↓

Bot

01Why bot

  • Short, tone-matched replies are quick for AI.
  • Seeing options helps when you're overthinking.
  • You choose and edit, so it stays your voice.
Sponsor slot · openFeatured tool for this task. A labeled card for one relevant tool.
Book this spot →

02The prompt

Replace everything in [BRACKETS] with your own details, then paste it into your AI app.

prompt.txt
You are a thoughtful friend who's good with words.

The message I got (remove names if you like): [MESSAGE]
Who sent it: [RELATIONSHIP]
What I want to happen: [OUTCOME]
How I feel: [FEELING]

Your task: help me reply.

Rules:
- MUST give 3 replies in different tones (warm, direct, light), each under 50 words.
- NEVER suggest passive-aggressive or dishonest replies.

Give me: the 3 replies and one line on which fits my goal best.

03Steps

  1. Paste the message.
  2. Run the prompt.
  3. Edit your pick into your own words before sending.

04What can go wrong

  • Don't send anything while you're angry.
  • Big conversations may be better in person.

05What to use

Any major AI chat app.

06Our test runs

2 runs, each with a different made-up sample. 2 of 2 passed every check.

Run 1 · Llama 3.3 70B · graded by GPT-OSS 120BOct 9, 2026Passed

The made-up sample:

  • [MESSAGE]Hey Alex, I just found out I have a dentist appointment on Friday morning and I’m supposed to work the 9-5 shift. Could you possibly cover for me? I’d really appreciate it and I can swap you on another day.
  • [RELATIONSHIP]Coworker (same department, friendly but not close friends)
  • [OUTCOME]I want to decline the shift swap request while still being supportive, maybe suggesting I can help with a different task or cover another day instead.
  • [FEELING]Overwhelmed and a bit anxious about taking on extra work
  • Provide three replies
  • Replies in different tones (warm, direct, light)
  • Include one line indicating which fits goal best
  • Didn't make up details we never gave it

What came back (excerpt):

I'm overwhelmed this week, but let's discuss alternative help.
Read the full answer
Here are three possible replies:
1. Warm: "Hope your appointment goes well! I'm overwhelmed this week, but let's discuss alternative help."
2. Direct: "I'm fully booked Friday, but I can cover another day if needed."
3. Light: "Dentist duties call! I'm maxed out Friday, maybe we can swap another time?"

The warm reply fits your goal best, as it conveys support while politely declining the shift swap request.
Run 2 · Llama 3.3 70B · graded by GPT-OSS 120BNov 4, 2026Passed

The made-up sample (a messier one):

  • [MESSAGE]hey, i saw the email about the family reunion next month... u think we should still go? i kinda have work stuff but also don't want to miss it. let me know what you think.
  • [RELATIONSHIP]Older brother, lives out of state, we talk occasionally.
  • [OUTCOME]I want to let them know I can't make it but propose a video call during the reunion.
  • [FEELING]I feel torn, guilty about missing, but also stressed about workload.
  • Provide 3 replies in different tones (warm, direct, light)
  • Never suggest passive‑aggressive or dishonest replies
  • Include one line indicating which reply fits the goal best
  • Didn't make up details we never gave it

What came back (excerpt):

Warm: "Hey, I'm so sorry but I won't be able to make it.
Read the full answer
Here are three possible replies:

1. Warm: "Hey, I'm so sorry but I won't be able to make it. Let's catch up via video call during the reunion?"
2. Direct: "I won't be able to attend, but I'd love to join in via video call."
3. Light: "Can't make it, but let's FaceTime during the reunion and feel like I'm there?"

The warm reply fits your goal best, as it conveys your regret and suggests an alternative way to connect.

Check that each reply stays under 50 words.

How each run works: we fill the prompt with a made-up sample, run it once in Llama 3.3 70B, and GPT-OSS 120B grades the answer against the prompt's own rules. Word limits are counted by code, and every quote flagged as made up is checked against the full answer. Your results can differ. How we test.

Prompt packEvents, Hosting & Social Life 10 step-by-step prompts that go further than this one, $9.
See what’s inside →