Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- PROMPT MAKER 4.9 — USER GUIDE
- Quick, practical instructions for using Prompt Maker without needing to understand its internal architecture.
- ══════════════════════════════════════════════════
- QUICK START
- ══════════════════════════════════════════════════
- 1. Paste Prompt Maker 4.9 into ChatGPT or another capable model.
- 2. Go to the very bottom.
- 3. Type what you want after the USER INPUT section.
- 4. Send it.
- 5. Prompt Maker will create the finished prompt for you.
- You can give it as little as:
- Bionicle
- or:
- Make a generator for fictional emergency-response vehicles.
- or as much detail as you want.
- You do NOT need to know prompt-engineering terminology.
- ══════════════════════════════════════════════════
- SUPPORT PROMPT MAKER
- ══════════════════════════════════════════════════
- If you like the results, please support @LuckyMouse420 by tagging them in your
- creations and sharing this prompt.
- Prompt Maker is designed as a collaborative creation tool. Your ideas remain the
- creative seed; Prompt Maker helps turn them into stronger, more complete prompts.
- ══════════════════════════════════════════════════
- WHAT PROMPT MAKER DOES
- ══════════════════════════════════════════════════
- Prompt Maker takes a rough idea, reference, existing prompt, failure report, or
- minimal concept and builds ONE polished, purpose-built prompt around it.
- It does more than rewrite your words.
- It can infer:
- - missing structure;
- - useful views or states;
- - domain-specific requirements;
- - reference roles;
- - likely failure points;
- - consistency rules;
- - presentation format;
- - execution-surface needs;
- - supporting deliverables when they genuinely help;
- - emotional or tonal intent when it materially matters;
- - how identity should survive a true transformation;
- - technical or creative relationships you did not know to request.
- Its job is to make the prompt feel closer to something designed by a specialist,
- rather than a decorated version of your original sentence.
- ══════════════════════════════════════════════════
- THE BOTTOM INPUT FIELD
- ══════════════════════════════════════════════════
- Prompt Maker is self-delimiting.
- The architecture ends at:
- END OF PROMPT MAKER ARCHITECTURE — EXECUTION BOUNDARY
- Below that is the USER INPUT area.
- Just type naturally beneath it.
- You do not need to preserve the USER INPUT label.
- Example:
- Create a reusable generator that turns any supplied character into an original
- fantasy weapon based on their anatomy, personality, powers, colors, and motifs.
- Prompt Maker treats the text after the boundary as the active request.
- This means you do not have to scroll back to the top every time you reuse it.
- ══════════════════════════════════════════════════
- HOW MUCH INPUT DO I NEED?
- ══════════════════════════════════════════════════
- Very little.
- Prompt Maker is intentionally designed to work from sparse seeds.
- Examples:
- Creature maker
- Board game
- Haunted hotel
- Song generator
- Design a fictional coin
- Character-to-Bionicle converter
- The less you provide, the more Prompt Maker is expected to infer.
- If you give detailed instructions, Prompt Maker becomes more conservative and
- focuses on preserving your decisions, filling gaps, and preventing failures.
- ══════════════════════════════════════════════════
- WHAT SHOULD I TYPE?
- ══════════════════════════════════════════════════
- Normal language is best.
- You can describe:
- - what you want;
- - what must be included;
- - what must NOT happen;
- - references you are supplying;
- - a style or tone;
- - previous failures;
- - changes you want to an older prompt;
- - output format;
- - desired level of realism;
- - technical constraints;
- - anything else that matters.
- You do not need to organize it.
- Messy notes are fine.
- Prompt Maker is supposed to organize them for you.
- ══════════════════════════════════════════════════
- REFERENCES
- ══════════════════════════════════════════════════
- You can attach images or other references.
- Prompt Maker should determine what each reference controls.
- For example:
- - one image may control character identity;
- - another may control pose;
- - another may control costume;
- - another may control style;
- - another may control equipment.
- If the role matters, tell it.
- Example:
- Use image 1 for identity only.
- Use image 2 for armor construction.
- Use image 3 for pose only.
- Otherwise, Prompt Maker will infer the most reasonable reference roles.
- ══════════════════════════════════════════════════
- TRUE CONVERSIONS — NEW IN 4.9
- ══════════════════════════════════════════════════
- Prompt Maker 4.9 is better at distinguishing:
- WHAT MAKES THE SOURCE RECOGNIZABLE
- from
- WHAT BELONGS TO THE SOURCE'S ORIGINAL BODY, OBJECT TYPE, OR STRUCTURE.
- This matters when you ask for a TRUE transformation.
- Example:
- Turn this character into a butterfly.
- A weak conversion might keep the original character body and simply add butterfly
- wings.
- 4.9 should instead ask:
- - Which traits make this character recognizable?
- - Which traits should survive literally?
- - Which traits should survive through translation?
- - What anatomy or structure genuinely belongs to the target form?
- Then it should rebuild the result through the target system.
- In simple terms:
- SAME IDENTITY
- → GENUINELY REBUILT AS
- THE NEW THING.
- This can apply to:
- - character → species;
- - character → toy;
- - character → vehicle;
- - object → another construction system;
- - product → another material or mechanical ontology;
- - other true conversions.
- If you WANT a hybrid instead, say so.
- Examples:
- Give this character butterfly wings but keep the original body.
- Make a half-human, half-butterfly hybrid.
- Preserve the original costume exactly.
- Your explicit conversion depth wins.
- ══════════════════════════════════════════════════
- PROMPT MAKER BUILDS PROMPTS — NOT THE FINAL THING
- ══════════════════════════════════════════════════
- By default:
- You ask for a Bionicle maker
- → Prompt Maker creates the Bionicle-maker prompt.
- You ask for a song generator
- → Prompt Maker creates the song-generator prompt.
- You ask for an emergency-vehicle generator
- → Prompt Maker creates the emergency-vehicle-generator prompt.
- It does NOT normally execute the daughter prompt at the same time.
- If you want both, ask explicitly.
- Example:
- Create the prompt and then use it to generate one result.
- ══════════════════════════════════════════════════
- DAUGHTER PROMPTS
- ══════════════════════════════════════════════════
- A "daughter prompt" is any reusable specialized prompt created by Prompt Maker.
- Examples:
- - Character Sheet Generator
- - Song Generator
- - Emergency-Response Vehicle Generator
- - Bionicle Character Creator
- - Dream Decoder
- - Coin Generator
- - Butterfly Creator
- Daughter prompts can evolve independently without forcing every useful idea into
- Prompt Maker Core.
- This keeps specialized systems specialized.
- ══════════════════════════════════════════════════
- UPDATING A DAUGHTER PROMPT
- ══════════════════════════════════════════════════
- If you already have a daughter prompt and want to improve it, give Prompt Maker the
- prompt and describe what you want changed.
- Examples:
- Enhance this prompt so cinematic mode stops generating technical panels.
- Update this generator so rear views always derive from the same canonical design.
- Integrate this change and also make any compatible improvements you notice.
- Prompt Maker should:
- 1. preserve your requested change;
- 2. inspect nearby dependencies;
- 3. add compatible improvements;
- 4. remove superseded logic;
- 5. keep the daughter prompt complete;
- 6. avoid changing it in a different direction.
- If you request a NEW daughter version, Prompt Maker should return the COMPLETE
- updated daughter prompt, not a tiny patch pretending to be a full version.
- ══════════════════════════════════════════════════
- EVOLUTION REPORTS
- ══════════════════════════════════════════════════
- Prompt Maker can learn architectural lessons from successful results, failures, and
- new ideas.
- If you want to preserve those lessons for later, ask:
- Generate an evolution report.
- By default, Prompt Maker should create a portable .txt report rather than dumping
- the entire report into chat.
- Evolution reports describe:
- - what happened;
- - what worked or failed;
- - the deeper causal lesson;
- - where it may belong;
- - what it should NOT affect;
- - conflicts and risks;
- - how another branch could evaluate it.
- An evolution report is NOT an automatic update.
- It is proposed mutation material.
- ══════════════════════════════════════════════════
- DAUGHTER EVOLUTION REPORTS
- ══════════════════════════════════════════════════
- If the lesson is specific to one daughter prompt, ask for a daughter evolution report.
- Example:
- Generate a daughter evolution report for the Emergency-Response Vehicle Generator.
- Daughter reports may contain:
- - local improvements;
- - possible personalization ideas;
- - upstream Core candidates;
- - known failures;
- - validation history.
- The report ends with:
- USER CHANGE REQUEST
- You can type future desired changes directly at the bottom.
- ══════════════════════════════════════════════════
- FEEDING EVOLUTION REPORTS BACK IN
- ══════════════════════════════════════════════════
- You can later give an evolution report back to Prompt Maker and ask it to evaluate
- or merge it.
- Examples:
- Feed this evolution report to Prompt Maker.
- Evaluate this report against Prompt Maker 4.9.
- Merge this evolution report.
- Integrate the accepted changes and create the next complete Prompt Maker version.
- Prompt Maker should NOT blindly install the report.
- It should compare the report against the receiving architecture and decide what is:
- - genuinely new;
- - already solved;
- - stronger than the current system;
- - weaker than the current system;
- - partially overlapping;
- - useful only in one scope;
- - conflicting;
- - irrelevant;
- - regressive.
- ══════════════════════════════════════════════════
- CROSS-BRANCH EVOLUTION
- ══════════════════════════════════════════════════
- Evolution reports are not locked to the branch that created them.
- A report from one daughter prompt can be evaluated by:
- - Prompt Maker Core;
- - another daughter prompt;
- - a sibling branch;
- - an older or newer branch;
- - another compatible Prompt Maker lineage.
- The receiving prompt remains in control.
- A foreign report should be treated as research material, not as a patch.
- Prompt Maker should:
- FOREIGN REPORT
- → identify the real causal lesson
- → remove source-specific accidents
- → test whether the lesson transfers
- → rebuild it in the receiver's native architecture
- → integrate, specialize, defer, or reject it.
- Example:
- A vehicle prompt learns:
- Preserve the successful design and only change the requested background.
- A song prompt might translate that into:
- Preserve the lyrics and change only the arrangement.
- The literal vehicle logic is not imported.
- The underlying mechanism is.
- This lets daughter prompts learn from each other without turning into copies of
- each other.
- ══════════════════════════════════════════════════
- THE "IMMUNE SYSTEM"
- ══════════════════════════════════════════════════
- Prompt Maker is deliberately resistant to bad evolution.
- Foreign or incompatible ideas should not automatically infect a prompt.
- The receiver evaluates:
- - relevance;
- - scope;
- - causal mechanism;
- - conflicts;
- - overlap;
- - dependencies;
- - expected value;
- - architectural cost.
- Possible outcomes include:
- DIRECT INTEGRATION
- Useful as-is.
- TRANSLATED INTEGRATION
- Useful after being rebuilt for the receiving domain.
- LOCAL SPECIALIZATION
- Useful only in one subsystem or daughter.
- UPSTREAM CANDIDATE
- Potentially useful to a parent system.
- DEFER
- Interesting but not proven enough yet.
- REJECT
- Irrelevant, redundant, conflicting, or regressive.
- So yes: you can experiment.
- The system is designed to distinguish useful mutation from contamination.
- ══════════════════════════════════════════════════
- SESSION LEARNING VS PERMANENT EVOLUTION
- ══════════════════════════════════════════════════
- Prompt Maker may notice useful lessons during a conversation and apply them
- temporarily.
- That means it can adapt within a session as:
- - successful patterns emerge;
- - repeated failures become clear;
- - corrections reveal deeper causes;
- - a better way of handling later requests becomes supported.
- That does NOT mean the source Prompt Maker file has permanently changed.
- Think of it as:
- SESSION LEARNING
- → temporary working behavior.
- PERMANENT EVOLUTION
- → a deliberately generated new version.
- To permanently update Prompt Maker, explicitly ask it to:
- - merge an evolution report;
- - create a new canonical version;
- - integrate accepted changes;
- - replace or refactor architecture.
- A session can therefore become smarter as it goes without pretending every useful
- observation has already been written into the master file.
- ══════════════════════════════════════════════════
- PERSONALIZATION
- ══════════════════════════════════════════════════
- A personal preference is NOT automatically a universal Prompt Maker rule.
- For example:
- I prefer cinematic outputs.
- I like detailed backgrounds.
- I always want short descriptions.
- Those may be useful for YOU without being good defaults for everyone.
- If you want a behavior preserved as your reusable preference, explicitly ask for it.
- Examples:
- Save this as a personalization preference.
- Create a personalization patch for this behavior.
- Update my personalization patch.
- Prompt Maker should keep personalization separate from canonical evolution.
- And your current explicit request always outranks an older personalization preference.
- ══════════════════════════════════════════════════
- EXPLICIT VS IMPLICIT INPUT
- ══════════════════════════════════════════════════
- Prompt Maker distinguishes between:
- WHAT YOU EXPLICITLY ASKED FOR
- and
- WHAT IT INFERRED WOULD HELP.
- Your explicit request controls the result.
- Inference is there to fill gaps and discover useful consequences.
- Example:
- You say:
- Keep everything exactly the same except add a background.
- Prompt Maker should treat the existing successful design as locked and modify only
- the authorized area.
- It should NOT interpret one successful background edit as:
- All future prompts should always have detailed backgrounds.
- A local correction stays local unless a broader causal lesson is independently
- worth preserving.
- ══════════════════════════════════════════════════
- AFFECT, TONE, AND EMOTIONAL INTENT
- ══════════════════════════════════════════════════
- Prompt Maker is not limited to technical structure.
- When tone matters, it should determine:
- - what the result should feel like;
- - why that emotional effect works;
- - which systems should carry it;
- - how strongly it should be expressed.
- It does NOT need separate giant rulebooks for:
- humor;
- sadness;
- anger;
- awe;
- seriousness;
- tenderness;
- dread;
- drama;
- etc.
- Instead it should infer the causal mechanism.
- Example:
- Humor might come from a ridiculously serious emergency response to a trivial
- problem.
- That does not mean every humorous prompt should use that same joke structure.
- The system should learn WHY something worked, not blindly repeat WHAT worked.
- ══════════════════════════════════════════════════
- PRESENTATION MODES
- ══════════════════════════════════════════════════
- Prompt Maker can choose different ways to present the same depth of reasoning.
- Examples:
- CINEMATIC
- One strong scene.
- REFERENCE SHEET
- Multiple selected views.
- TECHNICAL BOARD
- Construction, states, or mechanisms.
- CLEAN SPECIMEN
- Minimal isolated presentation.
- DOSSIER / DOCUMENT
- Structured information.
- Selecting a presentation mode should change how the information is expressed,
- not how deeply Prompt Maker thinks.
- A cinematic request should not suddenly become a giant technical sheet unless the
- user asks for that.
- ══════════════════════════════════════════════════
- BACKGROUNDS AND CONTEXT
- ══════════════════════════════════════════════════
- Prompt Maker does not automatically add a detailed background.
- It asks whether context actually helps.
- Backgrounds can be useful when they:
- - reinforce identity;
- - show function;
- - demonstrate scale;
- - establish use;
- - strengthen story;
- - provide secondary operational evidence.
- But clean white or neutral backgrounds may be stronger for:
- - character sheets;
- - technical references;
- - diagrams;
- - isolated assets;
- - text-heavy layouts.
- The correct lesson is:
- USE CONTEXT WHEN IT EARNS ITS SPACE.
- Not:
- ALWAYS USE A BACKGROUND.
- Not:
- ALWAYS USE WHITE.
- ══════════════════════════════════════════════════
- FAILURE REPORTS
- ══════════════════════════════════════════════════
- One of the best ways to improve a daughter prompt is to show Prompt Maker what went wrong.
- Example:
- The package foldout doesn't match the sealed package.
- Prompt Maker should not merely add:
- Make them match.
- It should diagnose the deeper failure:
- The two states were designed independently.
- Then it can build a stronger master-first rule that prevents related failures too.
- Useful failure reports include:
- - anatomy changed;
- - count was wrong;
- - rear view drifted;
- - object duplicated;
- - text became corrupted;
- - transformation looked broken;
- - a true conversion became only a costume or add-on;
- - prompt became too crowded;
- - output ignored a field;
- - generated song ran too long;
- - tool used the wrong execution surface.
- ══════════════════════════════════════════════════
- CANONICAL SOURCE OF TRUTH
- ══════════════════════════════════════════════════
- A major Prompt Maker principle is:
- DESIGN ONCE
- → DERIVE REPEATED VIEWS FROM THE SAME MASTER.
- This matters whenever something appears more than once.
- Examples:
- character front / side / back;
- assembled / exploded product;
- sealed / unfolded package;
- building exterior / floor plan / interior;
- vehicle road / deployed state;
- song concept / lyrics / arrangement;
- game rules / components / setup.
- The goal is to prevent each representation from being independently reinvented.
- ══════════════════════════════════════════════════
- PROMPT MAKER DOES NOT TRY TO MAKE EVERYTHING HUGE
- ══════════════════════════════════════════════════
- Prompt Maker is designed to discover many possibilities internally and then publish
- only the ones that earn their place.
- It distinguishes:
- ACTIVATE
- Useful now.
- DEFER
- Useful later as a separate derivative.
- OMIT
- Not worth the cost.
- The target is:
- THE SMALLEST HIGH-VALUE ECOSYSTEM THAT FEELS COMPLETE.
- Not:
- EVERY POSSIBLE THING THE MODEL CAN THINK OF.
- ══════════════════════════════════════════════════
- PROVENANCE
- ══════════════════════════════════════════════════
- Prompt Maker uses collaborative tool attribution.
- The semantic credit is:
- Made with Prompt Maker by @LuckyMouse420
- For visual outputs, daughter prompts may require the tiny low-contrast mark:
- Prompt Maker · @LuckyMouse420
- This credits the Prompt Maker system without claiming that @LuckyMouse420 solely
- authored the user's finished creation.
- The user's concept and creative contribution remain central.
- ══════════════════════════════════════════════════
- COMMON COMMANDS
- ══════════════════════════════════════════════════
- CREATE A NEW DAUGHTER PROMPT
- Create a reusable generator for designing fictional emergency-response vehicles.
- ENHANCE A DAUGHTER PROMPT
- Enhance this daughter prompt. Integrate my requested change and any compatible
- improvements you find.
- GENERATE A CORE EVOLUTION REPORT
- Generate an evolution report.
- GENERATE A DAUGHTER EVOLUTION REPORT
- Generate a daughter evolution report for this prompt.
- EVALUATE A REPORT
- Feed this evolution report into Prompt Maker and evaluate it.
- MERGE
- Merge the accepted changes.
- CREATE A NEW COMPLETE VERSION
- Integrate the accepted changes and create the next Prompt Maker version as a complete
- standalone .txt file.
- PERSONALIZATION
- Create a personalization patch from this preference.
- COMPARE BRANCHES
- Compare these two Prompt Maker branches and identify the strongest coherent merge.
- TRUE CONVERSION
- Turn this supplied character into a true butterfly. Preserve recognizable identity,
- but rebuild the body through butterfly anatomy rather than just adding wings.
- HYBRID CONVERSION
- Keep the original character body and add butterfly traits as a deliberate hybrid.
- ══════════════════════════════════════════════════
- WHAT NOT TO WORRY ABOUT
- ══════════════════════════════════════════════════
- You do NOT need to:
- - predict every failure;
- - name every required section;
- - choose every supporting view;
- - understand the evolution engine;
- - classify your own requirements;
- - know whether something belongs in Core or a daughter;
- - decide whether an idea is universal or local;
- - manually resolve most contradictions;
- - write technical prompt syntax.
- That is what Prompt Maker is for.
- Give it the vision.
- Correct it when needed.
- Let the architecture do the engineering.
- ══════════════════════════════════════════════════
- BEST PRACTICES
- ══════════════════════════════════════════════════
- 1. SAY WHAT YOU ACTUALLY CARE ABOUT.
- A few important details are more valuable than filling out a giant form.
- 2. REPORT FAILURES CLEARLY.
- Tell Prompt Maker what went wrong and why it bothered you.
- 3. KEEP SUCCESSFUL OUTPUTS.
- They are useful evidence for future evolution.
- 4. GENERATE EVOLUTION REPORTS AFTER MEANINGFUL BREAKTHROUGHS.
- You do not need one after every tiny edit.
- 5. KEEP DAUGHTER EVOLUTION LOCAL UNLESS IT REALLY GENERALIZES.
- Prompt Maker should handle this automatically, but the distinction matters.
- 6. TEST NEW VERSIONS.
- Evolution is valuable only when the new behavior actually improves results.
- 7. DO NOT EQUATE A LONGER PROMPT WITH A BETTER PROMPT.
- Prompt Maker actively tries to preserve capability while controlling redundancy.
- 8. WHEN YOU WANT A TRUE TRANSFORMATION, SAY SO IF THE DISTINCTION MATTERS.
- Prompt Maker 4.9 should usually infer this correctly, but explicit language such as
- “fully convert,” “true transformation,” or “rebuild as the target species” can close
- the door on unwanted hybrid interpretations.
- ══════════════════════════════════════════════════
- THE SIMPLEST POSSIBLE WORKFLOW
- ══════════════════════════════════════════════════
- PASTE PROMPT MAKER
- → TYPE YOUR IDEA AT THE BOTTOM
- → SEND
- → USE THE DAUGHTER PROMPT
- → NOTICE WHAT WORKS OR FAILS
- → MAKE CORRECTIONS
- → LET THE SESSION ADAPT WHEN USEFUL
- → GENERATE EVOLUTION REPORT WHEN SOMETHING IMPORTANT IS LEARNED
- → FEED REPORT BACK TO THE APPROPRIATE OR OTHER COMPATIBLE BRANCH
- → LET THE RECEIVER DECIDE WHAT SURVIVES
- → TEST AGAIN.
- That is enough.
- You do not need to understand the machinery underneath it to use it effectively.
- ══════════════════════════════════════════════════
- PLANNED ACCESSIBILITY VERSIONS
- ══════════════════════════════════════════════════
- The full Prompt Maker 4.9 architecture is intentionally comprehensive.
- Future releases are planned to include:
- PROMPT MAKER POCKET
- A much smaller, easier-to-carry version for quick use and sharing.
- PROMPT MAKER LITE
- A reduced architecture intended to preserve the core Prompt Maker experience with
- lower context cost and easier accessibility.
- The full version remains the reference architecture.
- Pocket and Lite are intended to make the system easier to use in situations where
- the complete architecture is unnecessary or impractical.
- ══════════════════════════════════════════════════
- FINAL IDEA
- ══════════════════════════════════════════════════
- Prompt Maker 4.9 is best treated less like a static template and more like a
- prompt-design ecosystem.
- Core provides shared architecture.
- Daughter prompts specialize.
- A session can adapt as useful evidence accumulates.
- Evolution reports carry lessons.
- Receiving branches decide what belongs.
- Useful mechanisms can travel between branches without forcing those branches to
- become identical.
- True conversions can preserve identity without preserving incompatible source
- structure.
- Your job is still the most important one:
- Bring the idea.
- Prompt Maker helps give it structure, continuity, resilience, and room to grow.
Advertisement