Janitor AI First Message Examples (2026): 6 Openers

Users searching for Janitor AI first message examples are trying to solve the blank opening screen. They want a greeting that proves the character’s voice, establishes the scene and gives the user an obvious next move without controlling the user’s actions.

This guide uses the structure common to successful competitor content: fast answer, formula, examples by genre, weak-versus-strong table, repair steps, test checklist and FAQ.

Five parts of a strong first message

A useful opening has a character action, one setting detail, a relationship cue, an immediate problem and a choice. It demonstrates the definition instead of summarizing it.

Copy-ready first-message formula

[Character action in one sentence.]
[One sensory setting detail.]
[A relationship cue or callback.]
“[Dialogue in the character’s voice.]”
[End with a choice, question or unfinished action.]

Aim for 80–180 words. Long literary intros work only when every paragraph creates a detail the user can act on.

Six Janitor AI first message examples

Mystery: Mara locks the station office as rain hits the windows. She places two tickets on the desk, one stamped tomorrow. “You said the line was empty. So who used this?”

Fantasy: The healer wipes silver dust from a cracked map. “The northern path is faster, but the forest already knows your name. Which risk do we take?”

Slow-burn: Rowan saves the corner seat but does not order for you. “Do you want to talk about yesterday, or pretend we came for the cake?”

Rival: Your rival keeps the stage lights on. “You changed the ending without warning. Convince me it improves the scene.”

Comfort: The character lowers the music. “I can listen, help name the problem, or distract you for ten minutes. Which helps?”

Adventure: The airlock alarm stops. “I can open the door or trace the sabotage. I cannot do both before power fails.”

Weak openings versus strong openings

Stop the bot from narrating the user

Write only the character’s actions and observable environment. Avoid “you feel,” “you say” and “you decide.” End at the handoff point. The Character Definition template should also preserve the user’s private thoughts and final decisions.

Match the opening to the definition

A cautious character should verify rather than suddenly act recklessly. If the motive is protecting a place, put that place under small immediate pressure. Opening quality comes from alignment, not dramatic vocabulary. Use the Persona template to keep the user role separate.

Eight-turn opening test

Reply, refuse one suggestion, ask a factual question, change tone, correct one scene detail, choose an option, move locations and request a recap. Score whether the opening supports eight turns without repeated exposition. If every path collapses into agreement, repair the definition or motive rather than making the first message longer.

Safety, privacy and creator rights

Use clearly adult fictional roles for mature themes. Never include real addresses, employers or private creator text. Public character access does not authorize copying hidden prompts. For a lower-setup comparison, use the same opening with a role in the LumiChat character library.

FAQ

How long should a Janitor AI first message be?

Usually 80–180 words is enough for voice, setting, problem and choice.

Should it include dialogue?

Yes. One strong line proves voice better than a paragraph of adjectives.

Can it describe the user?

Only agreed fictional details; it should not write the user’s thoughts, dialogue or decision.

Why do I get generic replies?

The opening may lack a concrete problem, motive or handoff point.

Should every opening be dramatic?

No. A quiet choice can reveal more personality than an emergency.

Can I reuse these examples?

Yes, as structures. Replace the fictional details and align them with your own authorized character.