Versioned method
Every result displays seed, engine aurora-1.0.0, rules phonotactics-en-fictional-1.0.0, and applied tags. The safety and recognizable-name filter is safety-ip-1.0.0.
Field Desk 01 · People
Character Name Generator
Create 12 reproducible fictional names with pronunciation, phonotactic notes, rule versions, and conflict checks against your world.
PRIVATE PREVIEW · NOINDEX- Output
- 12 distinct candidates
- Method
- Deterministic phonotactics
- Data route
- Browser only by default
- Release
- Blocked from public indexing
Generated folio
Candidate names
Ready to generate locally.
Your naming folio is empty
Leave the brief blank for a safe default, or add one line of world context. Generate is the only step before copying your first name.
Evidence contract
A generator you can inspect.
Reproducible tests
Fixtures cover same-seed reproduction, 12-result uniqueness, conflicting constraints, Unicode sanitization, filters, Board migration, and malformed imports. Test date: 20 August 2026.
Known boundaries
Outputs are fictional suggestions, not cultural etymology, identity validation, trademark screening, copyright clearance, or a promise of uniqueness.
Maintainer status
Product and engineering maintenance: Manus AI, preview implementation. A named human domain reviewer remains required before public production release.
What this tool returns
The interactive tool appears before this explanation so a search visitor can complete the task immediately. Its server-rendered description remains readable if JavaScript is unavailable.
| Example candidate | Why it fits | What to inspect |
|---|---|---|
| Vaelorin | Balanced consonants with a legible stress path. | Pronunciation, tags, seed, rule version, and Board conflicts. |
| Mireth Saren | A contrasting middle cluster separates it from nearby names. | Pronunciation, tags, seed, rule version, and Board conflicts. |
| Coravel | A compact ending supports related demonyms or family lines. | Pronunciation, tags, seed, rule version, and Board conflicts. |
How the local method works
The engine selects from a fictional phoneme inventory, applies weighted word positions and C(V)C syllable templates, then enforces length, consonant-cluster, spelling, prohibited-pattern, duplicate, and recognizable-fiction filters. It does not splice a database of real personal names and does not call an AI API.
Similarity warnings compare exact normalized text, shared openings, edit distance, approximate consonant skeletons, and syllable frames. These signals are writing aids only. Review names with appropriate cultural, editorial, and legal expertise before publication.
Standards and sources
The implementation is documented against named primary standards rather than unsupported claims. Unicode normalization supports stable text comparison; IndexedDB provides the browser-local record store used by the Board; WCAG defines the keyboard requirement tested by the acceptance suite.
“Unicode Normalization Forms are formally defined normalizations of Unicode strings which make it possible to determine whether any two Unicode strings are equivalent to each other.”
“All functionality of the content is operable through a keyboard interface without requiring specific timings for individual keystrokes.”
For local persistence behavior, see the W3C Indexed Database API 3.0. These sources describe platform behavior; they do not endorse generated names or guarantee legal uniqueness.