Home  /  How it works

The method · in full · nothing hidden

How a Names Done Right report is actually made.

Short version: the facts in your report — every popularity rank, every numerology number, every nakshatra syllable, every meaning flag — are computed by code from real data. AI assists with candidate expansion and report synthesis, while deterministic data and safety checks validate the output before its scheduled reveal. Here is every step, with nothing glossed over.

The dataset behind every rank
0 distinct names

Unique names in the U.S. Social Security Administration baby-names dataset for the 2025 data year — the public record our popularity engine looks each name up in, rather than guessing.

SourcePublic SSA dataset
Engine reads itIn code · no AI
Every figurePinned to its data year
We areAn independent tool
Code computes. AI assists. Automated checks validate.

That separation is the whole point. A name report is only worth paying for if its facts are true. So we never let a language model invent a rank, a meaning, a syllable, or a chart position — those come from deterministic engines and real datasets. AI can assist with candidate expansion and turn grounded inputs into clear, warm, personal writing; deterministic audits check the report against those inputs. We think that distinction is what separates a real report from a chat transcript.

The pipeline at a glance

Five stages, in order, every time

Nothing here is improvised. Each report runs the same path — and you can see exactly which steps are deterministic code, which use AI, and which need you.

Stage 1
Guided intake

Your taste, constraints, surname, and any opt-in tradition — structured, with unbundled consent.

You + form
Stage 2
Deterministic engines

Code computes the applicable facts: popularity data, safety-screen results, and any tradition calculations you selected.

Code · no AI
Stage 3
Grounded synthesis

AI can assist with candidate expansion and report wording, grounded in the verified facts above.

AI · grounded
Stage 4
Data checks & QA

Deterministic checks verify the data-backed facts, supplied constraints, and output guardrails before scheduled delivery.

Automated checks
Stage 5
The reveal

A crafted, scheduled delivery — and a free post-birth nakshatra refinement if you chose that layer.

Delivered
Needs you Deterministic code (no AI) AI, grounded in computed facts
Stage by stage, in depth

Now the long version

This is the part most products won't show you. Here's what actually happens, with a real example of what each engine produces.

1Guided intake

We learn your family before we suggest a single name

A great name fits a specific family, so we start by understanding yours. The intake is structured — not a freeform chat — so every candidate later gets measured against the same set of things you told us.

  • Your surname (for the say-it-out-loud flow test) and any sibling names already chosen.
  • Sounds, lengths, and styles you love — and your hard no-gos.
  • Pre-birth-first: your due date, plus optional parents' birth details, so we can prepare ahead and refine later.
  • Whether you want a tradition woven in — opted into separately, never bundled or assumed.
Unbundled consent Edit or delete anytime Used only for your report
Birth dates, times, and places are sensitive. You give them only if you opt into a tradition that needs them, with a separate checkbox — and they're used only to compute your report. Full detail is in our Privacy Policy.
Sample intake · illustrative
What we capture, structured
SurnameBennett
Sibling already namedAria
Lovessoft vowels, 2 syllables"nothing in the top 10"
No-goshard-to-spell, family names
Due dateNov 2026
Tradition layerNumerology — opted in ✓astrology / nakshatra: left off
2Deterministic engines · no AI here

The facts are computed in code, from real data

This is the engine room, and it is the reason the report is trustworthy. Candidate scoring, popularity, and safety functions run in deterministic code; optional tradition engines run only when you select them. "Deterministic" means the same inputs always produce the same outputs — there is no model guessing, no improvisation, and nothing invented. The same name and relevant inputs will compute the same results every single time.

  • Popularity engine — looks up each name's real rank, count, and multi-year trend from the public U.S. SSA dataset (and international sources), pinned to a data year.
  • Numerology engine — when selected, applies the classic letter-value methods (Pythagorean and Chaldean) as fixed arithmetic — addition and digit-reduction, nothing more.
  • Nakshatra engine — when selected, computes the Moon's ecliptic longitude from open astronomical ephemeris data, then maps it to the nakshatra and pada to find the traditional starting syllables.
  • Language-safety screen — checks pronounceability, spelling friction, initials/monogram, and known unfortunate meanings or slang across major languages.
Reproducible Auditable No model involved
Engine output · sample: "Maya"
Popularity engine
U.S. rank (SSA, 2025)#62
Births that year3,704
5-year trend▲ gently rising
Numerology engine — Pythagorean
# fixed letter values, summed, then reduced M=4 A=1 Y=7 A=1 sum = 4+1+7+1 = 13 reduce = 1+3 = 4 # expression number
A computed number is a fact about the spelling, not a prediction. What a tradition associates with that number is presented as cultural meaning — never as destiny, luck, or a guaranteed outcome.
3AI synthesis · grounded

AI assists the report — grounded in structured facts

Here's exactly where AI comes in, and where it doesn't. The deterministic pipeline builds and scores the shortlist from structured intake and reference data. Where AI is used, it can assist with bounded candidate expansion and report wording; factual report elements come from the computed dossiers, not from model guesses.

The workflow instructs the synthesis layer to use the provided facts, show uncertainty where the data is uncertain, and avoid inventing details. Deterministic audits then check the report against those inputs before delivery.

Cites computed facts Can't overwrite a number Uncertainty is shown
If a name's meaning is genuinely debated, the report says it's debated rather than picking the prettiest version. We give you the honest reading, not a nicer-sounding one.
Grounded write-up · sample
"Maya" — why it fits you

"Maya pairs cleanly with Bennett — two soft syllables into a crisp surname — and sits beside your daughter Aria without rhyming or clashing. It's well-loved but not over-exposed: #62 and gently rising, so it's familiar without being a top-10 name."

Every bold fact traces toStage 2 engines
Meaning"water / illusion"roots vary by language — flagged
Invented claimsnone
4Data checks & QA

Automated checks run before delivery

Before delivery, the workflow runs fixed data and safety checks against source-backed popularity facts, the constraints you supplied, and output guardrails for clearer, safer treatment.

The shortlist is matched back to your intake — your no-gos, sibling set, and the heritage you asked us to honor — using the pipeline's structured criteria.

Every report, no exceptions Fixed automated checks
The automated check set
What the workflow verifies
Facts match the engines✓ checked
No invented meanings or stats✓ checked
Your no-gos respected✓ checked
Surname flow + sibling fit✓ checked
Traditions framed as culture✓ checked
Writing is grounded in approved inputs✓ checked
5The reveal

Delivered as a crafted reveal — not a raw file dump

Naming a person is a moment, so we treat the delivery like one. Your report arrives as a designed, scheduled reveal you can read on your phone, print, or gift — not a raw file the second you finish typing.

  • A beautifully formatted report — read it, save it as a PDF, print it, or hand it to a grandparent-to-be.
  • On the partner tiers, you and your partner each privately rank the shortlist on your own link, and we reconcile the two into the names you both rated highly.
  • Free post-birth nakshatra refinement: if you opted into the Vedic layer pre-birth, send the actual birth time afterward and we re-run that engine at no charge.
One free revision 14-day refund window
Honest timing
Why it isn't instant

Two reasons, both honest: the report completes its data checks (Stage 4), and a thoughtful reveal beats a file that lands mid-form-fill. Most reports are scheduled within a few hours of a completed intake, and we show you the planned reveal time before delivery.

Engines + draftautomated
Data checksbefore delivery
Typical turnarounda few hours
Deliveryscheduled reveal
Post-birth refinementfree, if opted in
The honest line on AI

What the AI does — and what it is never allowed to do

We use AI, and we'll always tell you so. The trust is in the boundary: a clean line between what's computed and what's written.

What AI does

  • Reads the verified facts from the deterministic engines and selects the strongest candidates for your stated taste.
  • Writes the warm, specific "why it fits you" — citing the computed rank, flow, and sibling fit.
  • Orders and frames the shortlist so it reads as a considered recommendation, not a dump.
  • Phrases tradition associations in plain, respectful, culturally-credited language.
  • Says "this is uncertain" or "sources differ" when the underlying fact is genuinely unsettled.

What AI is never allowed to do

  • Invent or alter a popularity rank, count, or trend — those come only from the data engine.
  • Make up a numerology number or a nakshatra syllable — those are computed arithmetic and astronomy.
  • Pick a prettier "meaning" than the sources support, or hide that a meaning is debated.
  • Promise luck, fortune, health, wealth, success, or any guaranteed outcome from a name.
  • Present a tradition as science, prediction, or medical / financial / legal advice.

If you've ever asked a free chatbot for baby names, you know the failure mode: a confident, fluent answer with no idea whether the popularity number is real. We removed that failure mode by design — the numbers can't come from the writer. They come from the engines, and the workflow validates the report against them.

The fair question

"Isn't this just ChatGPT with a nice font?"

It's the right thing to ask, and you deserve a straight answer. Here's the honest difference, point by point.

Where do the numbers come from?
A chatbot generates a plausible-sounding number from text patterns. We look it up in the public SSA dataset and compute the trend in code. Ours is reproducible and pinned to a data year; a chatbot's can change between two asks of the same question.
Can it make something up?
A chatbot will happily produce a confident meaning whether or not it's correct. In our pipeline the writer is structurally prevented from supplying facts — it only quotes what the engines computed, and flags what's uncertain.
What validates it before delivery?
A chatbot answer is whatever the model returned, instantly. Every Names Done Right report runs through fixed deterministic data and safety checks before it reaches you.
Can it settle a disagreement?
A chatbot can't have you each privately rank a shortlist and fairly reconcile the two into one shared list — surfacing the names you both love and the bridge names, without taking a side. Our partner tiebreaker does — that's structure and process, not prompting.
What about the traditions?
A chatbot improvises astrology and numerology from memory. We compute numerology arithmetic and the nakshatra from astronomical ephemeris data — and frame the result as respected culture, never as science or a promise of luck.
Show your sources

The data and methods behind every figure

We'd rather over-explain than ask you to take it on faith. Here's what feeds the engines.

U.S. name popularity

The official U.S. Social Security Administration baby-names dataset — the national record of how often each name was registered, going back to 1880 — drives every rank, count, and trend we report.

Public dataset · figures pinned to their data year · we are an independent tool, not affiliated with or endorsed by any government agency.

International popularity

For families outside the U.S. or naming across cultures, we draw on published national and international name-frequency sources to ground popularity beyond the U.S. picture.

Coverage varies by country; where data is thin, we say so rather than imply false precision.

Numerology arithmetic

The classic Pythagorean and Chaldean letter-value systems, applied as fixed arithmetic — sum the letter values, reduce to a single digit. The math is identical every time.

Presented as a respected tradition. A number describes a spelling; it is not a prediction or a promise.

Astronomical computation

For the Vedic nakshatra layer, we compute the Moon's ecliptic longitude from open astronomical ephemeris data, then map it to the nakshatra and pada to derive the traditional starting syllables.

Real astronomy, used in service of a cultural naming practice — credited to its lineage, never sold as fortune-telling.

Language & safety screen

Each candidate is checked for pronounceability, spelling friction, monogram/initials, and known unfortunate meanings or slang across major languages — the traps that are easy to miss until it's too late.

Reference lists are maintained and expanded over time; no screen is exhaustive, and we say so.

Meanings & origins

Commonly-cited etymologies and meanings from standard naming references, with explicit flags wherever a meaning is debated, varies by language, or is genuinely uncertain.

When sources disagree, the report tells you they disagree — it doesn't pick the prettiest one.

A note we mean: everything above describes how the product is built and how each report is produced — the engines, the sources, and the checks behind every figure. We won't dress that up with invented reviews, ratings, or numbers. What we can show you is exactly how it works, and a full sample report you can read end to end.

Optional layers · opt-in · respected as culture

How the traditions are handled

Most families never switch one on, and the core report is complete without any. If you do want one, here's the method — and the line we never cross.

Numerology

Each name's number worked out by the classic letter-value methods, with what that number is traditionally associated with — and how it sits with your family's.

Method: fixed Pythagorean & Chaldean arithmetic, computed in code. A respected tradition — never a prediction, never a promise of luck or success.

Western astrology

With a birth date, time, and place, we read the sun / moon / rising and suggest names whose feel and sound align with that chart's character.

Method: chart positions computed from open ephemeris data. A "chart-aligned vibe," offered as culture — never a horoscope of destiny.

Vedic nakshatra naming

The traditional path: birth date, time, and place → the Moon's nakshatra and pada → the auspicious starting syllables — so the shortlist honors a practice families have used for generations.

Method: Moon longitude from ephemeris → nakshatra / pada → syllable map. Named and credited to the 27-nakshatra tradition — never whitewashed, never sold as fortune.

Hebrew & biblical

Gematria name-values and the biblical / Hebrew roots and resonances of each name, for families who want that lineage reflected.

Method: traditional gematria values, computed; sources cited. Offered respectfully, as heritage — not as science or guaranteed outcome.
The line we never cross. Every tradition here is treated as a respected cultural and spiritual practice — named properly and credited to its lineage. None of it is science. We never promise luck, fortune, health, wealth, or success from a name, and nothing here is medical, financial, or legal advice. You choose any, all, or none — and you choose the name.
Standards we hold every report to

The checks every report runs

Same checks, every report. This is the quality bar, written down.

Facts trace to a source

Every rank, number, and syllable in the report can be traced back to the engine and dataset that produced it. Nothing free-floating.

Uncertainty is shown, not hidden

Debated meanings and thin data are flagged in plain language. We're clear about what's genuinely uncertain instead of sounding falsely sure.

Your constraints are honored

Your no-gos, sibling set, surname flow, and any heritage you asked us to honor are checked against the final shortlist.

No promise we can't keep

No "perfect name guaranteed," no luck or fortune, no medical / financial / legal claim. The choice stays yours; we make it informed.

If it doesn't land, we fix it — or refund you

Want different finalists, a different style, or a tradition dropped? You get one free revision to make it right. And if it's genuinely not for you, there's a 14-day refund window. We'd rather earn the keep than argue you out of a refund you want.

Who's behind this

No fake team. A named founder is reachable by email.

Names Done Right is a new, independent product from the studio Underline. We're not pretending to be a 50-person company — we're a small operation that built a careful tool, with a named founder and support email you can use. As we grow, this page grows with what's actually true. You'll never read an invented statistic here.

Checkout secured by Stripe One-time payment No card kept on file

We never see or store your full card number — payments are handled entirely by Stripe.

Now you've seen the method.
See it in a real report.

The fastest way to judge the depth is to read a full, labeled sample — built end-to-end the exact way yours would be. No email required.

One-time · no subscription · preview before you pay · one free revision · 14-day refund