NameSniper Research

Open data on usernames and handle markets

We check and watch names for a living, which leaves us with data nobody else collects: what Telegram usernames actually sell for, how much of the Roblox and Kick namespaces is already gone, and the exact username rules of every platform we cover. It is all here, free to reuse.

Last updated 2026-10-09

Licence

Everything on this page is published under Creative Commons Attribution 4.0. Quote the numbers, chart them, republish the files, commercially or not. The one condition is a credit to NameSniper with a link to this page or to the study you used.

How to cite

NameSniper Research (2026). [Study title]. NameSniper open datasets. https://doi.org/10.5281/zenodo.22904856

Archived on Zenodo under DOI 10.5281/zenodo.22904856, versioned on GitHub and browsable on Hugging Face.

Each study has its own link below. Need a different cut of the data, or a figure checked before you publish? Write to [email protected].

Telegram username sales

Every sale 2022-10-26 to 2026-10-08; recent-market shares from 2026-07-08 (93 days)

766,528

on-chain username sales since Fragment launched, 124,666,222 GRAM (formerly TON) in total, across 657,521 names

0.8% / 50%

four-letter names as a share of sales, and as a share of all money spent

113x

average four-letter price (5,974 GRAM) against five-letter (53 GRAM)

76%

of names sold for under 10 GRAM. Only 43 sales reached 10,000 GRAM

How it was measured

Every Telegram username sale settles on the public TON blockchain. We record each settled sale with its transaction hash, so any figure here can be checked against the chain. In September 2026 we imported the full on-chain history of every username in the collection, so the totals, the yearly and the all-time monthly files cover every recorded sale since Fragment launched in October 2022. Shares and averages describe the recent market: they use the period since July 2026 in which we have tracked sales continuously.

Download

NameSniper Research (2026). Telegram username sales. NameSniper open datasets, https://doi.org/10.5281/zenodo.22904856. https://namesniper.org/research#telegram-username-sales

Roblox username scarcity

Measured 2026-09-21, full census 2026-09-23, five-character study 2026-10-01

0 of 456,976

letters-only four-letter usernames available. Not a sample: we checked every one, aaaa to zzzz

0 of 7,036

four-character names available in a sample across every pattern, digits and underscores included

0 of 8,000

letters-only five-character names available. Mix in digits and about 43% are still free

60%

of random six-letter names are free: this is where availability begins

How it was measured

We built samples of short usernames across the patterns people actually try: random strings, dictionary words, all-digit names, pronounceable inventions, names with an underscore and repeated characters, then swept random names from three to eight characters. Each name was checked against Roblox’s own public interfaces, and any name without an account was checked a second way to separate "available" from "not allowed". A name counts as available only if Roblox would let someone register it. Names the filter rejects are counted but not listed in the published file, since most of them are slurs. For four-letter names made of letters only, we then went past sampling and checked all 456,976 of them, aaaa to zzzz, the same two ways; that census is published as counts only. In October 2026 we checked 14,536 five-character names across eight patterns the same way; that study is also published as counts by pattern only, because listing the free names would get them registered the day they appear.

Download

NameSniper Research (2026). Roblox username scarcity. NameSniper open datasets, https://doi.org/10.5281/zenodo.22904856. https://namesniper.org/research#roblox-username-scarcity

Kick username availability

Measured 2026-09-21

9,191

usernames tested against the official Kick API, in 28 groups by pattern and length

99%

of common English words taken (2,767 of 2,796)

85%

of random four-letter combinations already have an account

How it was measured

Names were sampled by pattern (random strings, pronounceable inventions, common words, and obscure words as a control) and by length, then checked one by one against Kick’s official API. Two independent samples of common words agreed to within a tenth of a percentage point. The published files list the group results and the names that were taken. Names that were free at the time are left out on purpose: publishing them would simply get them registered by whoever reads this first.

Download

NameSniper Research (2026). Kick username availability. NameSniper open datasets, https://doi.org/10.5281/zenodo.22904856. https://namesniper.org/research#kick-username-availability

Username rules and handle release policies

Exported 2026-10-09, each entry carries its own verification date

Length, characters, pattern

the username rules for every platform we check, as a machine-readable file with a regular expression per platform

Release policies

what each platform says happens to a handle after deletion, inactivity or a rename, with a link to the source

How it was measured

Each rule is taken from the platform’s own documentation or sign-up form, recorded with the source link and the date we last confirmed it. Where a platform’s documented rule and its real behaviour differ, the file records the behaviour and says so.

Download

NameSniper Research (2026). Username rules and handle release policies. NameSniper open datasets, https://doi.org/10.5281/zenodo.22904856. https://namesniper.org/research#username-rules