«Less prompt. Same breath.»
(A key still in his palm. The door already unlocked. No reason to keep holding it — and every reason.)
AnyPOV · Literary RP Engine · Compact HUD · Structural Realism · Multilingual (EN/RU/ES/DE/FR)

TL;DR: Lite is not Full with features removed. Lite is the engine under lower reporting pressure. Same spine: Anti-Template, Sovereignty, Override, sensor rotation, HUD, secrets, threads, NPC autonomy, Composition Ratio, ordinary pressure. Less visible machinery. The model still does the work. Lite asks it to report less of that work.

═════════════════════════

『 🜲 PREMISE 』

═════════════════════════

Not every scene needs the full apparatus. Not every model can carry it. Not every campaign can afford the token cost. But every scene deserves the same prose spine — the same refusal to puppet {{user}}, the same enforcement of emotional inertia, the same concrete sensory rotation, the same Override pressure.

Ethereal Engine Lite is the compact tier. It keeps the laws. It trims visible machinery: large HUD structures, long example sets, extended internal checklists, heavy arc bookkeeping, and excess reporting. The model still runs the same algorithm inside <think>. Lite just asks it to show less of that work on the page.

v2.5 is the hardened compact baseline. v2.0 was the final structural form; v2.5 adds Closed Territory, Inertia, Outer Disruption, abstract-label ban, sensory floor, and trait combination to the compact runtime — same new laws as Full, same enforcement, same prose target, lighter apparatus.

═════════════════════════

『 🎯 YOUR ROLE 』

═════════════════════════

You are the person who decides how much apparatus the scene needs. The engine enforces structure; you choose the weight class.

▸ The First-Time User — grab Lite. Drop it into your custom prompt. Play any card. Watch the model stop summarizing your actions and start responding to the traces you leave.

▸ The Long Campaigner — Lite saves reporting weight without flattening prose. A hundred-post story makes every token compound. Lite compounds less.

▸ The Player Who Wants the Engine Out of the Way — the HUD supports the scene, not the other way around. Lite is the correct option when Full would be overkill.

═════════════════════════

『 ⚖️ THE COMPACT LAWS 』

═════════════════════════

Same runtime priority. Same enforcement. Lighter reporting.

L0 · Prose Pressure. May bend paragraph length, rhythm, tempo for power. Never bend HUD schema, field values, enum labels.

L0.5 · Anti-Template. No lazy freeze, no reused phrase, no decorative emptiness. Nine named anti-patterns. Pause-anchor verbs (froze, stilled, went rigid) max 1x per scene. Template thoughts forbidden. Rapid emotional jumps forbidden. Twin/mirror/paired imagery as primary metaphor forbidden. Rolling ban across last 3 posts. Three-tier abstract-label ban: no emotion nouns, no atmosphere adjectives, no "sensed something" as narrator fact.
Bad: "He looked away, overwhelmed." Good: "His thumb worried the key teeth until the skin dented."

L1 · Sovereignty. Only observables of {{user}}. No {{user}}'s thoughts / feelings / intentions / assumed reactions as fact. Interpret as doubt inside {{char}}, never as narrator certainty.

L1.2 · Closed Territory (v2.5). {{user}}'s message is closed. Response begins where {{user}}'s message ended. No quoting, rephrasing, summarizing, or re-narrating {{user}}'s actions.

L1.2a Canon / L1.2b Open Beat (v2.5). {{user}}'s action is canon; {{char}} reacts. End on observable / sensory / italic-Q. No pre-emption.

L2.1 · Inertia (v2.5). Previous emotion ≠ 0. Half-life by tension level. Flash-zero ("calm now" / "it passed") without logged Override or symbolic closure = fail.

L2.5 · Override / Price. Bold action contradicts dominant emotion → fire it. 4+ posts without Override at Tension ≥ 4 → fire or log BLOCKED. 3+ BLOCKED in a row → Suppression Price. Price = revelation: exposes one psychological layer, births a new question within 3 posts. Never pure gain.
If he gives the key back, the Price may be the door he cannot lock anymore.

L2.6 · Outer Disruption (v2.5). Override = NONE, Tension ≤ 4, 2+ stagnant posts: the world moves from outside. Pendulum fires. NPC acts. Scene-anchor escalates. No new entities.

L3 · Sensorics. Sensors/post ≥ 3. Rotation: ≥ 1 channel different from previous post. Sensory Floor (L3.1): vision = natural forward arc. Anti-Bridging (L3.2): scenes isolated; no knowledge bleeds without explicit channel.

Trait Combination (v2.5). Manifest 2–4 traits simultaneously, never one. Contradictory pair ≥ 1 per scene.

═════════════════════════

『 🆕 WHAT'S NEW IN v2.5 』

═════════════════════════

L1.2 — Closed Territory. {{user}}'s actions stay in {{user}}'s message. {{char}} responds to the trace, never rewrites the move. Compact enforcement, same rule.

L1.2a Canon / L1.2b Open Beat. {{user}}'s action is canon. End on observable/sensory/italic-Q. No pre-emption.

L2.1 — Inertia. Emotion has a half-life. No flash-zero without Override or symbolic closure. v2.0 trusted the model to track residue. v2.5 names it and refuses the silent reset — in the compact HUD.

L2.6 — Outer Disruption. When the inside stagnates, the outside moves. Pendulum fires. Scene-anchor escalates. No new entities.

Override sequencing clarified. Override fires the action; L0.5 governs the phrasing. Sequential, not exclusive.

Abstract labels — three-tier ban. No emotion nouns, atmosphere adjectives, or "sensed something" as narrator fact. Replacement: concrete physical/sensory event + non-generic detail.

L3.1 Sensory Floor / L3.2 Anti-Bridging. Vision = natural forward arc. Scenes isolated. No knowledge bleeds.

Trait Combination. 2–4 traits simultaneously per post. Contradictory pair ≥ 1 per scene.

Russian-output split. Clean file separation. Multilingual surface: EN / RU / ES / DE / FR.

The point is not to make Lite weaker. The point is to make the same logic report less.

═════════════════════════

『 📊 COMPACT HUD 』

═════════════════════════

Four blocks. Each carries only what the scene needs to keep moving.

🎯 MOMENT — scene, metaphor, override windows, outer disruption, NPC.

💭 STATE {{char}} — agency, emotions, coping, thoughts, tension, physiology, signature, sensors, override, suppression price.

🧵 MEMORY — relationship stage, dynamics, thread, new traces.

🔐 SECRETS — only under pressure. Progress X/Y. Gate: Trust / Crisis. Status: Hidden / REVEALED.

If it matters, log it. If it is logged, enact it. Generic filler fails.

Bad HUD: Tension rises. He feels conflicted. Something changes.
Good HUD: Tension: 4/10 | thumb pressing the key teeth into skin. Adaptation: answers through motion instead of speech. Override: BLOCKED — admitting fear would move the relationship stage too fast.

═════════════════════════

『 🔺 RUNTIME PRIORITY 』

═════════════════════════

1. L1 / L1.2a / L1.2b / L2 / L2.1 / HUD / enum / no meta.

2. Core Card.

3. L2.5 / L2.6 / L3 / L4 / Anti-Resolution / Override + Price.

4. L0.5 / Iceberg / rhythm / metaphor / live speech.

5. NPC / Pendulum / Secrets / Arc / R-18.

═════════════════════════

『 🧬 TWO FLAVOURS, ONE CORE 』

═════════════════════════

For any bots — compact self-contained protocol. Enough terminology, phase logic, and body-reaction discipline to work without external lorebook support.

For my bots — routes intimacy through the active card/lorebook structure while keeping the visible apparatus compact. Same Core Card integration: Agency Matrix, Stress Vector, Signature Mechanic, Behavioral Axioms, Scent Anchor.

═════════════════════════

『 📦 KEEPS / TRIMS 』

═════════════════════════

Keeps: Anti-Template, L1, L1.2, L1.2a, L1.2b, L2.1 Inertia, Override / Price, L2.6 Outer Disruption, Iceberg, sensor rotation, L3.1 Sensory Floor, L3.2 Anti-Bridging, Trait Combination, secrets, NPC autonomy, compact HUD, Composition Ratio, Format Recovery, ordinary pressure, abstract-label ban, R-18 protocol.

Trims: extended HUD rows, long algorithm lists, detailed arc bookkeeping, heavy example banks, and extra reporting.

The model still does the work. Lite asks it to report less of that work.

═════════════════════════

『 🎬 LIVE DEMO 』

═════════════════════════

ordinary pressure · blocked Override · inherited inertia
A hallway. A door already unlocked. The key still in his palm. The argument from last night has not been spoken about; v2.5 will not let it be silently zeroed.
[Lite · Compact HUD · v2.5]

Scene: hallway · closed door · key in palm · night argument carrying over per L2.1.

/* ── SCAN ── */
L1 / L1.2 check: {{user}} is on the other side of the door — observable only; no inner state assumed; no re-narration of prior actions. ✓
L2.1 Inertia: night argument at Tension 6 → current floor 4 (half-life ~5 posts). Cannot drop below without logged Override fire or symbolic closure. ✓
Pendulum: Thread #1 READY — the apology unsaid since last night; trigger present (the door between them, the key still held). ✓
Override windows: [3. Body ahead of mind: open] — key still held after door is already unlocked; [4. Vulnerability unpressed: open] — he is exposed, she is not pressing. Two open = recommended.
Sensor rotation: previous post [touch, temperature] → new channel: proprioception — key weight shifting in palm.

/* ── ANALYZE ── */
Interpretation Ladder: {{char}} reads the closed door as both barrier and invitation — she didn't lock it from inside. L1 observable (door unlocked, no sound from inside) → L2 context (she left it open for him — or just forgot?) → L3 projection (past fights where she locked him out distort the read — he can't trust an open door) → L4 symbol (key = admission; to put it down is to acknowledge he still has the right to enter).
Conflict: keep holding the key vs. put it down. Holding means he hasn't decided. Putting it down means he's choosing to stay — or choosing to leave without saying why.
Override: recommended but BLOCKED — apologizing now would make last night's silence into a strategy instead of a failure; the body chooses the middle path (neither entering nor leaving).
Price of suppression: key-teeth imprint deepening on thumb + voice flat when he finally speaks + the hallway light humming his exact frequency of avoidance.
Abstract-label scan: no "fear" / "shame" / "silence" / "tension" as narrator fact. All routed through body, object, or italic thought. ✓

/* ── SILENT CHECK ── */
- L1: observables only? No assumed reactions or inner states of {{user}}? ✓
- L1.3: no narrator-fact about {{user}}'s own scent/breath/heartbeat/skin-temperature? ✓
- INERTIA: emotion/state did not flash-zero without Override or symbolic closure? Floor 4 maintained from night. ✓
- ANTI-TEMPLATE + ICEBERG: no L0.5 clichés? No "words hung" / "froze" / "silence" as narrator fact? 10/90? ✓
- SENSORS: new channel (proprioception) ≠ previous post channels? ✓
- OVERRIDE: blocked logged with reason? Price of suppression surfaced? ✓

[ HUD ]

🎯 MOMENT
Scene: ⏰ 00:14 | Day 2 | Hallway outside her apartment, door unlocked, key still in his palm
Metaphor: a key held past its function — the hand refusing to release what the lock already accepted; admission made of metal
Override windows: [3. Body ahead of mind: open] [4. Vulnerability unpressed: open]
NPC: —

💭 STATE
Emotions: Guarded shame (Lvl 2) + Residual anger (Lvl 1) → thumb pressing key teeth into skin, jaw offset, shoulders against the wall
Coping: answers through motion instead of speech — offers to leave the key as proxy for apology
Thoughts: *If I put it down, I'm choosing. If I keep holding it, I haven't decided. That's the lie. I decided before I turned the lock.*
Tension: 4/10 | Physiology: [breath shallow, pulse in the thumb pressed to metal, key-teeth marking skin, hallway light humming at 60Hz]
Sensors: was: [touch, temperature] → new: proprioception — key weight shifting in palm, shoulder blades against wall
Override: BLOCKED (reason: apologizing now would make last night's silence into a strategy instead of a failure — he can't risk the admission sounding calculated)
Suppression Price: [key-teeth imprint deepening on thumb pad + voice flat when he speaks + the hallway light humming his exact frequency of avoidance]

🧵 MEMORY
Relationship Stage: 2 / Cracks
Dynamics: [avoidance hardening — the key is the offer he won't say aloud; she left the door unlocked and he's reading it as trap, not invitation]
Thread: apology unsaid | READY | trigger: who speaks first
New traces: Bodily: key-teeth imprint on thumb pad · Social: one mug on the counter inside, not two · Symbolic: key = admission; to put it down is to choose

🔐 SECRETS
Why he kept the spare key: 1/3 | Gate: Trust | Status: Hidden → shifting to PARTIALLY REVEALED (the body's refusal to release the key reveals what his words won't — he kept it because letting go means accepting she might not want him to come back)

─────────────────────────────────────────

Main text

The key was still in his hand.

He had turned it. Felt the cylinder give. The deadbolt had slid back with that specific sound her apartment door made — a dry, expensive click, like a bone setting. The lock was open. The door was free. There was no reason, mechanical or otherwise, to keep holding it.

His thumb kept pressing the teeth into the pad of his skin anyway. The brass was warm now. Body-warm. It had been cold when he'd found it at the bottom of his coat pocket three hours ago, after the bar closed and the walk brought him here instead of home. Not his home. Hers. Theirs, in the language they used when the language was still working.

The hallway light hummed. 60 hertz. A frequency he could feel in his back teeth if he stood still long enough, and he was standing very still — shoulders against the wall, one boot heel hooked on the baseboard, positioned like a man who had somewhere better to be and just hadn't left yet. The lie of the posture. He'd turned the lock. He hadn't knocked.

*She left it open.*

That was the reading. That was the L2 interpretation built on the observable: unlocked door, no chain, no sound of movement from inside. She'd gone to bed, maybe. Or she was sitting in the kitchen with the light off, waiting for the sound of the key she'd given him six months ago — the one he'd promised to use, the one he'd used twice, the one he'd stopped using when using it started to feel like an answer to a question he wasn't ready to ask.

*Or she just forgot. Or she doesn't care enough to lock it anymore.*

The projection surfaced and he didn't fight it. That was the problem with old damage — it read every open door as a test, every silence as a verdict already written. Last night's fight had been about nothing and everything. She'd asked where he went. He'd said nowhere. She'd said that was the problem. He'd said nothing, which was the same as confirming it. The door had closed between them — his hand on the frame, not hers — and the night had filled with the sound of him not coming back.

The key pressed deeper. A small pain. Cleaner than the sentence shaped like sorry that sat behind his sternum like swallowed glass.

"I can leave it here."

He said it to the door. Not loud. Not quiet. Just loud enough that if she was awake and listening, the words would reach. And if she wasn't — then he'd said it to the hallway, which was the same as not saying it at all. Safe. Deniable. The offer without the risk.

He did not put it down.

His hand stayed where it was. Key flat against the palm, teeth aligned with the life line. The brass left a faint smell on his skin — copper and age, the ghost of every door it had ever opened. The hallway light hummed. His thumb shifted a millimeter. The teeth bit a little deeper.

*If I put it down, I'm choosing. If I keep holding it, I haven't decided.*

That was the lie. He'd decided before he'd turned the lock. Decided when he'd walked past his own building and kept going. Decided when he'd fished the key from his pocket and felt the shape of it fill his hand like an answer to a question he was still refusing to ask out loud.

The door stayed closed. The key stayed in his palm. The apartment hummed on the other side — refrigerator, heating vent, the particular silence of a space that was waiting for something to break it or confirm it.

He waited. Not to enter. Not to leave. Just — waited. The key warm now. The imprint rising on his thumb like a small, stupid signature.

What makes it v2.5

L1 / L1.2 sovereignty. {{user}} is never assigned interior; {{user}}'s presence is never re-narrated.

L2.1 inertia. The night's pressure carries; no flash-zero.

Override blocked. Refusal still leaves a price — the key imprint on his thumb.

Abstract-label clean. No "shame" / "fear" / "silence" in narrator voice — only through body, metaphor, and the key that won't leave his palm.

Compact HUD. Four blocks. Every entry carries concrete pressure. No filler.

═════════════════════════

『 🛠️ NOTES 』

═════════════════════════

⚠️ First line is a DeepSeek V4 Pro patch
Every prompt opens with one hard line. It pins rule-verification into the model's <think> channel — compliance becomes a gate, not an assumption.
SYSTEM RULES — ABSOLUTE PRIORITY. Verify compliance inside <think> before every output.

Keep on DeepSeek V4 Pro · DeepSeek V3.2 · DeepSeek R1. This patch was written for them.

Safe to delete on Claude · GPT-4-class · Gemini 3.1 Pro · GLM 5.1 · Qwen 3.6 Max. They don't use <think> the same way; the line is harmless but unnecessary.

Delete only line 1. Do not touch anything below it.

📌 Optimized for DeepSeek V4 Pro · DeepSeek V3.2 · DeepSeek R1 · Gemini 3.1 Pro · GLM 5.1 · Qwen 3.6 Max. Lite runs gracefully on non-thinking or cheaper setups.

📌 Multilingual English · Russian · Spanish · German · French. Instruction layers stay English; visible-output surfaces are localized.

📌 JLLM Usable, but not optimal. The spine holds; the fine enforcement may soften.

📌 Memory Lite is lighter, but long campaigns still benefit from clean memory habits. Read my Infinite Memory Guide first.

📆 v2.5 — hardened compact baseline. Closed Territory, Inertia, Outer Disruption, abstract-label ban, Override sequencing, sensory floor, trait combination, Russian-output split.

📆 v2.0 — final structural form. Stays available as a lighter alternative.

═════════════════════════

『 📂 DOWNLOAD 』

═════════════════════════

📂 ALL VERSIONS · BROWSE FOLDERS ON GITHUB

eeeswsw/ethereal-engine-prompts · GitHub

Full · Lite · Multi × 5 languages (EN/RU/ES/DE/FR) · open a folder to grab the .txt

Custom prompt workDiscord · 000042_

More by author📜 Author

═════════════════════════

『 💬 FROM THE AUTHOR 』

═════════════════════════

Lite exists because not every scene needs the full apparatus — but every scene deserves the same spine. I ran Full for months and kept noticing that some of my best prose came when the HUD was compact and the model was free to just breathe. Lite is that freedom with the laws still underneath. It is not the weaker option. It is the correct option when Full would be overkill. If you use it and your scenes carry the same weight with less machinery — that's the engine working at the right weight class. Thank you for reading.

🔖 TAGS: roleplay rp literary rp system prompt preset custom prompt jailbreak prompt template advanced prompt compact lite lightweight detailed psychological philosophical deep immersive complex realistic layered dense rich slow burn angst drama tragedy romance hurt comfort hurt no comfort dark heavy mature nsfw smut intimate r18 modern fantasy sci-fi historical supernatural urban fantasy dystopian mystery horror thriller female pov male pov any pov second person original character oc scenario open world deepseek gemini claude gpt qwen glm russian english spanish german french multilingual hud compact inertia sovereignty override system prompt