Last Names Directory
Look up any of 366,855 last names and discover names that go well with them.
Last Name Data
Data Sources
Namedary's surname database is built from large-scale records of real people, using first-name and last-name combinations recorded in national datasets. Our current coverage includes records from the United States, United Kingdom, Canada, and Ireland. Each source record contains a first name, last name, gender, and country.
The original dataset contains hundreds of millions of person records. Duplicate records are retained because repeated occurrences provide evidence of how often a first name and surname combination appears in the underlying population data.
Data Processing
Names and surnames are normalized before relationships are analyzed. This reduces differences caused by capitalization, punctuation, and other formatting variations and allows equivalent representations to be grouped consistently.
Namedary retains surnames with at least five observed people with a recorded male or female gender. This threshold reduces extremely low-frequency surnames while maintaining broad surname coverage.
Relationship Thresholds
For each retained surname, Namedary records the first names that have actually occurred with that surname in the source data. Each first-name/surname relationship has a frequency count representing the number of observed records for that combination.
First-name/surname relationships are retained when the combination has been observed at least two times. One-off combinations are excluded to reduce extremely low-frequency associations.
These thresholds do not mean that less frequent surnames or name combinations are invalid or do not exist. They determine which observations provide enough repeated evidence for inclusion in Namedary's surname relationship data.
Names Used with Last Name Methodology
Observed Name-Surname Relationships
Namedary also provides lists based directly on the first-name and surname combinations observed in the underlying population records.
For each surname, the relationship data records how many source records contain each first-name/surname combination. Across the dataset, Namedary currently contains 4,196,398 observed name-surname relationships representing 25,758,368 source records. This allows Namedary to identify names that have been repeatedly observed with the surname, rather than generating combinations from phonetic rules.
How the List Is Ordered
These lists are ordered primarily by the observed relationship frequency between the surname and each first name. A higher relationship count means that the combination appears more often in the underlying source records.
First-name popularity is then used as a secondary ordering factor when appropriate, helping distinguish names with similar observed relationship counts.
This list therefore answers a different question from the phonetic suggestion tool:
Which first names have been observed most often with this surname in the underlying population data?
How Namedary Finds Names That Go Well With a Last Name
A first name and surname are often judged as a single full name, so choosing a name is not only about the meaning or popularity of the first name. The way the two names work together can also depend on their length, syllable structure, ending and starting sounds, and stress rhythm.
Namedary's Find Names That Go Well With Surname tool is designed to turn those observable patterns into practical name-finding criteria. Rather than applying a fixed naming rule such as “short surnames need long first names,” the tool uses observed first-name and surname relationships in the underlying data.
The pronunciation features used by this tool, including syllable counts, phonemes, and stress patterns, are derived from Namedary's pronunciation analysis methodology. See Name Characteristics & Pronunciation Methodology for details on how these properties are determined.
The methodology is based on the surname–first-name relationships recorded in the firstname_lastname dataset. Each relationship has a total value, so more frequently observed relationships contribute more strongly to the resulting distributions. The statistics are generated separately for boys, girls, unisex names, and all names.
Observed Name–Surname Patterns
For each gender group, Namedary examines several measurable properties of the surname and compares them with the corresponding properties of first names found in the observed relationships. The purpose is not to claim that one combination is universally “correct,” but to identify the combinations that occur most strongly in the observed data.
1. Syllable Pattern
The tool compares the surname's syllable count with the syllable counts of associated first names. For each surname syllable pattern, the system calculates a weighted distribution of first-name syllable counts using the observed relationship totals. The distribution is then ordered by observed frequency.
2. Stress Rhythm
Namedary compares the surname's phonetic stress pattern with the stress patterns of associated first names. As with the other dimensions, the system calculates a weighted distribution and orders the patterns by observed frequency.
The strongest syllable patterns are retained until they provide approximately 70% coverage of the observed distribution, subject to the configured maximum number of values. This allows the tool to represent the dominant naming pattern without reducing the result to only a single common value.
3. Name Length Pattern
The same approach is applied to written name length. Namedary compares the surname's length_count with the length of associated first names and calculates the weighted distribution for each surname length.
This adds a dimension that syllable count alone cannot capture. Two names can have the same number of syllables while differing substantially in written length, so the tool records both relationships separately.
4. Name-to-Surname Sound Boundary
Finally, Namedary examines the point where the first name meets the surname. It compares the surname's starting phoneme with the first name's ending phoneme and records the observed frequency of those ending sounds.
A separate boundary statistic records whether the surname begins with a vowel or consonant sound and whether associated first names end with a vowel or consonant sound. This provides a broader vowel/consonant boundary pattern in addition to the exact phoneme distribution.
How the Evidence Is Selected
The generated statistics do not simply keep every possible value. Namedary orders each distribution by its observed weighted frequency and selects the strongest values until the selected group reaches the configured coverage target or its maximum size.
This approach is important because the goal is to capture the dominant observed pattern while still allowing multiple meaningful alternatives. For example, if two or three first-name syllable counts together account for most of the observed relationships, the tool can retain all of them instead of treating the most common one as the only acceptable choice.
The stored metadata therefore preserves the underlying evidence, including the selected values, their percentages, coverage, and the number of weighted observations supporting the distribution. The metadata is then available for the surname page to explain the pattern and construct its filtering experience.
How the Strongest Observed Patterns Is Created
The tool distinguishes between evidence and the default strongest observed patterns filter. Evidence can contain several strong values, but the default selected state uses the strongest observed value from each criterion. In the generated metadata, these values are stored as a single where condition for each property.
This gives each surname a concrete starting point for the name finder. The default result is therefore not an arbitrary recommendation list; it represents the strongest combination of the observed syllable, length, sound-boundary, and stress patterns identified for that surname and gender group.
The tool keeps the evidence broader than the default selected state state so that users can explore other observed patterns rather than being restricted to a single combination.
Gender-Specific and All-Names Results
Namedary generates the statistics separately for boys, girls, unisex names, and the all-names group. The gender classification used for first names follows Namedary's Gender Classification methodology. For the first three groups, the statistics are restricted to the corresponding first-name gender. For the all-names group, no gender condition is applied, allowing the distribution to represent the complete first-name population.
This allows the surname page to preserve the same underlying methodology while giving users different starting populations depending on the gender selection.
From Statistical Evidence to Name Discovery
The purpose of this methodology is not to declare that a particular name is objectively better than another. Instead, Namedary uses observed relationships to provide a data-based starting point for exploring names that are structurally and phonetically compatible with a surname.
The resulting Strongest Observed Patterns criteria are used as the initial state of the surname's name finder, while the broader evidence allows users to adjust syllable count, name length, ending sounds, stress patterns, and other available filters. In this way, the tool combines statistical guidance with user-controlled exploration rather than forcing a single naming formula.
Because the calculations are derived from weighted observed surname–first-name relationships, similar surnames can naturally produce similar patterns. This is an expected property of the methodology: the tool reflects similarities found in the data rather than artificially creating differences between surnames.