Random Name Generator
Generate random first and last names for characters, testing or inspiration, with control over origin and gender association.
Eight letters never start a name here
The lists behind the button are the fifty commonest American male first names, the fifty commonest female ones, and the fifty commonest surnames — 5,000 possible full names, which sounds like plenty. It is also frequency-ranked, and frequency-ranked lists are not spread evenly across the alphabet:
| List | Initials used | Never appear |
|---|---|---|
| first names | 18 of 26 | F I O Q U X Y Z |
| surnames | 18 of 26 | E I O Q U V X Z |
So the generator reaches 324 of the 676 possible pairs of initials: 47.9%. There is no Fiona, no Oscar, no Yolanda, no Ursula, and no surname beginning with E or V. If you are generating placeholder people and sorting them by initial, half the alphabet is empty — and it is the same half every time you press the button.
And the letters it does use are lumpy
| Initial | First names | Share | Against an even spread |
|---|---|---|---|
| J | 14 of 100 | 14% | 2.52× |
| S | 11 of 100 | 11% | 1.98× |
| A | 10 of 100 | 10% | 1.80× |
| M | 8 of 100 | 8% | 1.44× |
| D | 7 of 100 | 7% | 1.26× |
| R | 7 of 100 | 7% | 1.26× |
J starts 14 of the 100 first names — James, John, Joseph, Jennifer, Jessica and nine more. An even spread across the 18 available initials would be 5.6% each, so J runs at 2.5 times its share, and the top three initials between them account for 35% of every first name the tool can produce. Surnames are flatter: the commonest initial there is M, with only 5 of 50.
None of that is a mistake in the lists. Real names are distributed like this, which is exactly why the top fifty look like this. The mistake would be assuming a uniform draw from a frequency-ranked list gives you a realistic population — it gives every name equal weight, so the output ends up simultaneously too narrow in its alphabet and too flat in its frequencies. Two different distortions, pulling opposite ways, from one shortcut.
One small thing worth knowing: Thomas and Scott appear in both the first-name and surname lists, so Thomas Thomas and Scott Scott are both reachable. With 5,000 combinations that will happen about once in every 2,500 names.
How to use
- Choose the style or cultural origin you want.
- Generate as many names as you need at once.
- Regenerate freely — nothing is saved between runs.
- Search any name before publishing it.
- Treat every result as a starting point, not a decision.
Frequently asked questions
What is this actually useful for?
Naming fictional characters, populating test data, seeding a demonstration, and getting past the blank-page problem when an unnamed character is holding up the writing. Any generated name is a prompt rather than a decision.
Could a generated name belong to a real person?
Almost certainly, since the components are real names in common use. That is unavoidable and normally harmless — but it is worth searching any name you intend to publish, particularly for a fictional villain or in anything commercial, because coincidences do occasionally cause real problems.
Are the names culturally accurate?
Approximately, and no more than that. Name lists reflect broadly common usage rather than the full picture, and combining a first and last name at random can produce pairings that would be unusual or incongruous in practice. For fiction set in a specific culture, someone who knows it beats any generator.
Why do names go in and out of fashion?
Because naming is strongly cyclical — names tend to feel dated for several decades and then become available again, which is why many names popular now were common four generations back. It is one of the more reliable patterns in social data, and it is why period fiction so often gets names subtly wrong.
Should I avoid names close to real public figures?
For anything published, yes. Fiction that appears to depict a real person, particularly unfavourably, creates a genuine legal risk, and the closer the name and surrounding circumstances the greater it becomes. Distinctiveness is the safer direction to err in.
Does anything get uploaded?
No. Names are assembled in your browser from local lists, so nothing is transmitted and no account is needed.
Why do some combinations sound wrong?
Because first and last names carry their own regional and generational associations, and pairing them at random ignores that. A name that sounds implausible usually has a mismatch of era or origin behind it, and regenerating is faster than reasoning about which.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.