One brief per tool
Each of the 208 tools has an English label, two four-part word banks, an original design rule, and an invented story starting point. There are four families: characters, worlds, objects, and groups or stories. A harbor and a personal name do not use the same final layout even when both draw on familiar images. The catalog data and generator code can be inspected directly.
What a seed controls
The tool, seed, tone, format, and optional motif determine a reproducible ordering of combinations. Asking for six results instead of three preserves the first three. A fresh seed changes the order; the same seed and settings under unchanged data reproduce the same batch. The system avoids duplicates within a batch. It does not ensure global uniqueness, and another visitor can obtain the same name.
How settings change a result
Tone changes the pool of descriptive words. Compact format favors a short pairing; storybook format adds a genre-appropriate framing word or title. An optional motif replaces the first word, which lets a project reuse a chosen anchor. The interface constrains motif length to keep results manageable. The motif is plain text and is escaped when rendered.
Why a creative note is not a meaning
A result includes a short design interpretation and a scene exercise derived from the tool’s brief. These notes help you decide how to use a candidate; they do not report a real etymology, historical usage, or a universal symbolic meaning. The word combinations do not understand your full setting. Rewrite any option that becomes awkward when spoken or placed beside the rest of your cast.
Where AI assistance fits
AI assistance was used to develop this edition’s English briefs, prose, code, and main illustration. A visitor’s generator request does not invoke a live AI service. No secret external research dataset or verified linguistic corpus is claimed. The source makes the finite word banks and deterministic rules available for inspection.
Checks and limits
The release checks verify tool counts, metadata, local links, English document language, generation repeatability, batch uniqueness, PNG image references, and absence of embedded ad or analytics code. Those checks do not establish legal availability, cultural suitability in every language, independent expert review, accessibility certification, search ranking, or AdSense acceptance. Originality here describes newly developed briefs and examples, not a guarantee that no phrase has ever appeared elsewhere.
Subject-specific briefs rather than one interchangeable form
Every tool has three named directions and a distinct field label, professional brief, common pitfall, and practical test surfaces. An orc identity can focus on craft, family promise, or self-definition, while a city can focus on trade, a founding route, or a revised public map. Those choices introduce different anchors into the candidate process. The shared interface supports comparison, but the underlying questions and original scene continuation are specific to the subject.
Genre and sound are explicit rules
Genre supplies a small additional pool appropriate to an invented project frame. Sound preference reorders candidates using counts of selected visible English consonant patterns. That spelling heuristic does not transcribe pronunciation, model every language, or assign a professional quality score. A fixed seed controls reproducible variation among the candidates. Read the resulting names aloud and check actual neighboring labels rather than treating the option called flowing or crisp as a promise about every reader’s experience.
Names are assembled by category
Character identities can use a personal form, an epithet, or an alias; places can favor a landmark, compact settlement form, or formal geographic phrase. Objects and group or story titles have their own structures. Format controls separated words, joined compounds, or storybook presentation. Extended length can add a background-based phrase. These rules create alternatives for editing; they do not establish a real cultural naming convention, historical title, material property, or official identity.
Practical restrictions are applied before selection
The optional opening letter, main-name character cap, and comma-separated exclusion terms remove candidates that fail the selected restriction. The tool does not quietly ignore a limit in order to claim it generated the requested number. Short length also restricts the visible main form. If too few candidates remain, the page reports that result and suggests relaxing a condition. Formal and short alternatives are labeled separately from the main draft so their different communication jobs can be considered.
Ordering, distinct openings, and larger batches
The current engine builds a finite candidate set, removes duplicate names, and uses seeded ordering and visible heuristics to select results. It prefers different openings before returning further unique combinations where more are requested. Choosing three, six, or twelve does not change the underlying ordered set, so the earlier results remain the prefix of the larger batch under unchanged data and settings. That property supports controlled comparisons, not a guarantee of global uniqueness or inexhaustible novelty.
Five parts of a result
A result explains its construction, visible sound and shape checks, background and original scene, short and formal forms, and a revision pass. The scene and interpretations are fictional. The visible counts are ordinary properties of the text. They do not verify cultural suitability, audience recall, trademarks, accessibility, or a name’s real-language origin. The project-specific evidence still comes from your own work, and an explanation should be treated as an invitation to edit rather than an instruction to preserve every generated association.
Versioning and earlier saved entries
The expanded engine uses a version marker. Its settings and grammar differ from the earlier edition, so an older seed may produce a different batch. Existing notebook entries continue to show their saved text rather than being erased or silently rewritten. Export any exact record you need. The source catalog and rules remain inspectable, which helps explain how a result was assembled, but inspectability is not a blanket license to republish the entire collection under another identity.
What comparison and reports can establish
The comparison desk places up to four current drafts together and exposes their forms, counts, focus, and scene. Batch and notebook exports preserve a readable working record. A report link prepares relevant details in an email draft for your review. None of these features sends material to a live model, gathers public votes, or certifies the final creative choice. The purpose is to make decisions and limitations visible enough that a writer can revise deliberately and a correction can be investigated.