Begin with a moment, not a dictionary
Put the subject in a situation before choosing the longest or rarest word. A courier delivering a difficult letter needs a name another person can call from a doorway. A floating city needs a name residents can put on a ticket. An organization needs a name under which someone can ask it to do work. These practical uses expose weaknesses that a list of beautiful syllables can hide.
Read the opening scene on a topic page and choose one branch. Notice who is affected by the decision and who gets to describe it. The title a character claims may differ from the name a neighbor uses. That difference can become a source of story rather than an error to eliminate.
Separate the naming layers
For a person, distinguish a given name, a family or house association, an office, and an earned epithet. You do not need all four in every introduction. A close companion may use only the first; an official register may preserve the family; a hostile witness may repeat the epithet with a different meaning.
Places can likewise have official, local, and visitor names. Organizations can have a charter-name and a working abbreviation. Artifacts can have a maker’s designation and a later legend. The workshop’s formal and everyday structures help you compare these layers without pretending they are interchangeable.
Choose the naming community
Begin with the topic’s three root entries and their stated creative associations. The first community favors a single root with a world-register ending; the second links two roots; the third reverses their order for a ceremonial compound. Construction notes expose the actual spelling units, including contracted roots and merged boundary letters. These are invented conventions for a fictional world, not claims about an existing language.
World texture changes the joining vowel and register endings. Temperament changes the preferred root order and sound treatment. Standing changes the formal introduction: an emerging practitioner, an established resident, an elder or an office holder need not be addressed alike. Feminine, masculine and uninflected labels describe fictional ending styles; they do not determine the identity of the person you are creating.
Each result separates its short call name, formal introduction, optional alias and family or source lineage. Copy one form when you need a line of dialogue; copy the full brief when you need construction and context. A custom core has no inferred etymology: the page’s root association is offered separately, so your chosen word never acquires a fabricated translation.
When two good names fail together
Mira, Miren and Mirra can each look attractive in isolation. Put all three in a hurried conversation and the reader may have to stop to remember who spoke. A longer surname does not always solve the problem because close companions often use only a short form. Compare the form that actually appears in dialogue, not just the elaborate introduction at the start of a chapter.
The project check compares the current batch with up to fifty names you supply, one per line. It checks full and first-word spellings, repeated three-letter openings and one-letter differences. Case, punctuation and common accent marks are normalized. A warning is a reason to read the names together; it is not a command to reject them. Related people may intentionally share a sound, while different spellings can still be pronounced alike without triggering this check.
Reference names stay in the current tab. They are not included in the saved shortlist or exported candidate files, and the checker does not send them to a search engine or external service. It cannot detect every linguistic association, pronunciation conflict or existing fictional character. Keep sensitive manuscript material out of the reference list; a small list of fictional names is enough.
Test names in company
Read the candidate aloud in an ordinary sentence and beside the other names in the project. Three characters with the same opening syllable may be easy to confuse even if their complete names are different. Very long titles can work in ceremony but become cumbersome in fast dialogue.
The comparison table places short names, formal names and aliases side by side and reports core-name letter counts. Construction units show the root fragments and endings actually joined by the generator. These are spelling units, not a phonetic transcription or authoritative syllable count. Try the name in the language your audience will read, and choose your own pronunciation rather than claiming a certified Latin reading.
Use constraints deliberately
The sound control changes the pool of eligible given names or core words. The length control applies to the same part. A required initial further restricts that pool. A locked core word takes precedence over sound and length because you have explicitly chosen it, but exclusions still apply to every displayed name form, including the family or maker name.
Excluded fragments are comma-separated and case-insensitive. If you exclude “aur,” a displayed name containing that sequence is rejected even when it appears inside a longer family name. Very restrictive settings can produce fewer candidates or none. The tool reports that result rather than silently relaxing the rule.
A seed makes comparisons repeatable. Keep it fixed while changing one setting to understand that setting’s influence. Change it when you want another sample. A seed is not evidence that a name is unique, and a future engine revision may change the same seeded output.
Test the situation behind the name
Start with the topic’s three quantities and read the labels carefully. They have different meanings on different pages: an arrival window is not a stock interval, and a corroboration threshold is not a vote. Run the model and inspect the calculation detail. Every factor is visible, so you can disagree with an assumption instead of treating the result as an unexplained score.
The starting-assumption comparison reports numerical changes from the page’s original setup. Decision sensitivity changes one option at a time while keeping the others fixed. Quantity comparison varies one selected input above and below the current value. Inputs stay inside their bounds, and whole-unit quantities are rounded down; this can make two nearby perturbations produce the same result.
The feasible-plan search varies the capacity input named in its result, leaves the other quantities fixed, and tests combinations of only the decisions you permit. Its minimum is within the stated bounds and search increment. A plan can satisfy the model while being undesirable for your story, so inspect its choices before applying it. No hidden delivery, extra worker, missing consent, or additional vote is invented to rescue a failed run.
A composition plan shows coordinates and a reference square rather than producing an image. A permission model shows failed gates and cooldowns rather than forcing an activation. A queue can leave eligible people waiting. Use those limits as material for a scene. The worksheet preserves whether the current conditions hold. The naming controls above keep names and aliases separate from numerical scenario labels.
Scenario JSON includes the current inputs, assumptions, quantities, sensitivity rows, and perturbation results. Name JSON and CSV retain the selected scenario alongside the candidate and its story notes. Downloads are created in the browser. These are transparent fictional planning tools, not forecasts, professional advice, or certificates of historical accuracy.
Give the interpretation a source
A familiar word can carry a narrative association without explaining a personal name’s historical origin. The collection uses motif glosses to prompt choices: ash can mark an aftermath; a key can mark access with conditions; a bell can connect a warning to a listener. These are editorial readings, not proof of ancestry or grammatical correctness.
If you need an authentic historical identity, choose a time, location, language, and social context, then consult specialist sources. Do not assume that an ancient family name has a transparent modern meaning or that every Latin-inspired compound is an attested phrase. Keep creative notes separate from factual assertions in a published project.
Revise through dialogue and consequences
Write a short exchange in which somebody uses the name and somebody else responds. Does the form fit their relationship? Would they really say the title in this room? Does the reader need the entire name now, or would a familiar form make the scene more immediate?
Next, let the central choice produce a consequence. A guardian protecting a grove may have to hear from the village using its bridge. A healer may need a boundary before another promise. An artifact may work only under conditions its owner dislikes. A name earns emotional weight when the story gives other people reasons to remember it.
For portraits and visual briefs
Choose one material, one gesture, and one object that carries the scene. A repaired seam can explain more than a pile of unrelated symbols. Specify where light comes from and what it reveals; a reflective material needs different treatment from fabric or weathered wood.
The portrait brief is text you can revise for your own artwork or image tool. Lorafeed does not render an image or send your brief to an image service. Names, lighting cues, and material choices are prompts, not guarantees of what a separate image system will produce.
Keep a useful record
Save candidates only when you want persistent storage in the current browser. A shortlist can include several topics, so a name for a city can sit beside the organization that governs it. Open a saved entry to review its original forms, construction and notes, or copy its complete brief without regenerating the batch. Compare the exported record with your project notes and add the reason you chose one version.
Choose whether to export the current batch or the saved shortlist. Export text for a readable working document, JSON for structured reuse, or CSV for a spreadsheet. Saving a name does not silently change the export source. Treat exported files as your own responsibility. Clearing local storage removes this browser’s copy; it does not remove downloads or copies shared elsewhere.
Before publication
Check important names against the real-world and fictional contexts relevant to your project. Consider existing businesses, prominent characters, unintended meanings, pronunciation barriers, and the perspectives of people whose languages or traditions inspired the work. Commercial identity checks may require professional advice.
The workshop helps you make a deliberate creative choice. It does not perform a trademark search, certify originality against every existing work, verify a complete etymology, or guarantee commercial success. If you find a specific problem in the collection, the Contact page explains how to report it.