Statistics on the Gemini space

This page presents some statistics on the current state of the Gemini space. It has been updated on 2025-01-01 04:04:00Z.

It cannot claim to represent the entire space. The real number of URIs is certainly higher. There are several reasons why many URIs are not in the database:

the capsule may forbid retrieval, through robots.txt,
we do not know all the URIs and some cannot be found from the ones we know,
Lupa has a maximum number of URIs per capsule, to save resources (currently 10000).

On this page, "working" means there was a successful connection recently. "recently" means "less than 31 days". "Dead" URLs and capsules are removed after 46 days and no longer appear in any statistics.

Currently, our database includes 674,552 URIs, 528,923 of them having been checked successfully (status code 20) and recently. Among the recently accessed, 396,523 URIs serve a Gemini content.

Resources

The average size of the resources is 60,000 bytes.

Quantiles

10% of the resources are 266 bytes or less,
20% of the resources are 576 bytes or less,
30% of the resources are 1,012 bytes or less,
40% of the resources are 1,200 bytes or less,
50% of the resources are 2,051 bytes or less, MEDIAN
60% of the resources are 3,562 bytes or less,
70% of the resources are 6,111 bytes or less,
80% of the resources are 11,861 bytes or less,
90% of the resources are 74,057 bytes or less,
100% of the resources are 4,156,230 bytes or less.

# Quantiles only for Gemini pages

10% of the resources are 237 bytes or less,
20% of the resources are 446 bytes or less,
30% of the resources are 773 bytes or less,
40% of the resources are 1,089 bytes or less,
50% of the resources are 1,250 bytes or less, MEDIAN
60% of the resources are 2,104 bytes or less,
70% of the resources are 3,454 bytes or less,
80% of the resources are 5,659 bytes or less,
90% of the resources are 8,828 bytes or less,
100% of the resources are 4,156,230 bytes or less.

Ranges

Less than 10 bytes: 3849 URLs (0.73 %)
10 to 100 bytes: 14711 URLs (2.8 %)
100 to 1000 bytes: 138644 URLs (26.2 %)
1 to 10 kbytes: 256580 URLs (48.5 %)
10 to 100 kbytes: 67091 URLs (12.7 %)
100 to 1000 kbytes: 29155 URLs (5.5 %)
More than 1000 kbytes: 18893 URLs (3.57 %)

Most common media (MIME) types

text/gemini: 396,523 URLs
text/plain: 30,969 URLs
image/png: 25,562 URLs
image/jpeg: 23,412 URLs
application/pdf: 12,796 URLs
application/octet-stream: 10,150 URLs
text/x-go: 5,491 URLs
image/svg+xml: 3,116 URLs
application/zip: 2,215 URLs
image/gif: 2,046 URLs
audio/mpeg: 1,912 URLs
text/html: 1,374 URLs
text/xml: 1,286 URLs
octet/stream: 1,155 URLs
application/x-mscardfile: 1,081 URLs
MIME: 997 URLs
text/x-diff: 914 URLs
application/atom+xml: 696 URLs
application/javascript: 667 URLs
audio/ogg: 644 URLs

Most common languages

Unspecified: 419,174 URLs
en: 78,811 URLs
de: 11,267 URLs
it: 7,159 URLs
fr: 6,653 URLs
es: 1,869 URLs
es_ar: 1,273 URLs
ja: 586 URLs
ru: 560 URLs
en_gb: 415 URLs
en_au: 242 URLs
en_us: 143 URLs
pl: 110 URLs
sv: 109 URLs
eo: 80 URLs
gl: 55 URLs
pl_pl: 43 URLs
zh: 39 URLs
sco,gd,it,en: 38 URLs
uk: 37 URLs

Most common encodings ("charsets") for all files

(Remember there exists testing capsules, with very exotic encodings, so don't be surprised by some strange ones.)

Unspecified: 427,369 URLs
utf-8: 101,382 URLs
us-ascii: 100 URLs
binary: 46 URLs
gzip: 22 URLs
iso-8859-1: 2 URLs
bzip2: 2 URLs

Most common encodings for gemtext files only

Unspecified: 309,814 URLs
utf-8: 86,706 URLs
iso-8859-1: 2 URLs
us-ascii: 1 URLs

By the way, 335 of recently tested URLs (0.052 %) have a wrong encoding (it does not match the actual content).

Status codes

(Remember there are test capsules with funny status codes, to exercice Gemini clients.)

20 (Success): 528,923 occurrences (89.50 %)
51 (Not found): 26,054 occurrences (4.41 %)
40 (Temporary failure): 11,760 occurrences (1.99 %)
42 (CGI error): 10,782 occurrences (1.82 %)
60 (Client certificate request): 6,174 occurrences (1.04 %)
10 (Input request): 4,356 occurrences (0.74 %)
30 (Temporary redirect): 2,301 occurrences (0.39 %)
31 (Permanent redirect): 198 occurrences (0.03 %)
53 (Proxy request refused): 172 occurrences (0.03 %)
59 (Bad request): 80 occurrences (0.01 %)
50 (Permanent failure): 58 occurrences (0.01 %)
44 (Slow down): 46 occurrences (0.01 %)

Links

(We count only backlinks from external capsules, and at most one link per capsule. Also, we exclude links from capsules like search engines or directories.)

Maximum number of incoming links: 326

Average number of incoming links: 0.31

Capsules

There are 4246 capsules. We successfully connected recently to 2912 of them.

Most common capsules by number of working URLs

We have a limit of 10000 URLs per capsule.

midnight.pub: 10000 URLs
x5dragonfire.flounder.online: 10000 URLs
1436.ninja: 10000 URLs
mirrors.apple2.org.za: 10000 URLs
gemini.conman.org: 10000 URLs
gemini.techrights.org: 10000 URLs
fumble-around.mediocregopher.com: 9999 URLs
gemini.tuxmachines.org: 9999 URLs
dvd.flounder.online: 9997 URLs
gemlog.stargrave.org: 9997 URLs
gemini.oxydable.fr: 9996 URLs
taz.de: 9995 URLs
hoagie.space: 9993 URLs
station.martinrue.com: 9982 URLs
caiofior.pollux.casa: 9980 URLs
tjp.lol: 9956 URLs
library.inu.red: 9956 URLs
musicbrainz.uploadedlobster.com: 9950 URLs
oracular.space: 9949 URLs
gmi.noulin.net: 9948 URLs

Most common capsules by number of bytes in working URLs

We have a limit of bytes per URL.

Not properly documented yet

1436.ninja: 9588.6 megabytes
mirrors.apple2.org.za: 2818.6 megabytes
nytpu.com: 1993.1 megabytes
gem.librehacker.com: 1028.8 megabytes
librehacker.com: 979.2 megabytes
dfdn.info: 601.4 megabytes
yam655.com: 598.3 megabytes
hoagie.space: 495.8 megabytes
norayr.am: 494.1 megabytes
jpfox.fr: 403.6 megabytes
blitter.com: 374.0 megabytes
si3t.ch: 372.7 megabytes
b2khgkvb2wn4avjshjp63kknsjwikgwff5dwwydldia6qwf4kdnueyad.onion: 337.3 megabytes
tweek.zyxxyz.eu: 325.8 megabytes
library.inu.red: 322.8 megabytes
going-flying.com: 293.2 megabytes
ecs.d2evs.net: 276.4 megabytes
gael.mooo.com: 261.5 megabytes
canary.city: 239.2 megabytes
gemi.dev: 223.1 megabytes
shit.cx: 182.3 megabytes

All working capsules:

As a text file

As a gemtext, with links

Certificates

2654 (91.1 %) capsules are self-signed, 15 (0.5 %) use the Certificate Authority Let's Encrypt, 243 (8.3 %) are signed by another CA (may be not a trusted one).

87 capsules (3.00 %) have an expired certificate.

Algorithms:

ecdsa-with-SHA256: 1822 capsules
sha256WithRSAEncryption: 1016 capsules
ecdsa-with-SHA384: 48 capsules
ED25519: 21 capsules
sha512WithRSAEncryption: 5 capsules
ecdsa-with-SHA512: 3 capsules

Key types:

ECDSA: 1875 capsules
RSA: 1019 capsules
ED25519: 21 capsules

Key sizes for RSA:

2048: 746 capsules
4096: 263 capsules
3072: 7 capsules
1024: 3 capsules

Key sizes for ECDSA:

256: 1810 capsules
384: 64 capsules
521: 1 capsules

TLS

98 % of the capsules use TLS 1.3, 2 % use TLS 1.2.

robots.txt

313 (11 %) the capsules have a robots.txt exclusion file.

Ports

17 working capsules (0.6 %) use an alternative port

Addresses

1085 IP addresses used. 18 % are IPv6.

# Addresses with most virtual hosts

173.230.145.243: 914 vhosts
72.65.51.74: 484 vhosts
213.219.38.200: 282 vhosts
46.23.81.157: 124 vhosts
2a03:6000:1813:1337::157: 102 vhosts
90.65.170.44: 34 vhosts
109.237.26.252: 32 vhosts
45.56.93.217: 19 vhosts
128.140.115.191: 13 vhosts
212.71.248.87: 11 vhosts
2a03:6000:6f67:624::99: 10 vhosts
23.88.35.144: 10 vhosts
2a01:4f8:c17:20f1::42: 10 vhosts
46.23.94.99: 10 vhosts
51.222.161.16: 9 vhosts
143.244.212.63: 8 vhosts
2001:8b0:1d97:eeca:3da5:94ce:b61e:ee69: 8 vhosts
85.208.51.149: 8 vhosts
81.187.234.86: 8 vhosts
89.234.140.141: 7 vhosts

TLDs

There are 262 TLDs in the capsule's names, and 2045 registered domains.

Most common TLDs

# By number of registered domains

com: 322 domains
net: 182 domains
org: 172 domains
xyz: 145 domains
space: 96 domains
site: 65 domains
de: 61 domains
dev: 58 domains
me: 49 domains
eu: 38 domains
fr: 33 domains
uk: 33 domains
info: 27 domains
io: 26 domains
club: 23 domains
ru: 18 domains
online: 17 domains
onion: 17 domains
se: 16 domains
ca: 16 domains

# By number of capsules

(There's a strong bias towards TLDs which have hosting services such as flounder.online, which has many capsules in subdomains. See before the TLDs per registered domains, which are probably more useful.)