API
Free, keyless JSON for IRC networks and channels: find them, connect to them, and see how busy they are. Build it into your client, your bot or your site.
ircs:// link.
Stats
Live numbers and history
People online, rank, peaks and trends for any network or channel, with history from the last day to all time.
Members
Your network's private analytics
Run a network? Make a personal key and pull its peak hours, channel churn and daily history into your own tools.
Ready-made scripts
Don't want to write code? These add an IRC network finder to your client, using this API.
erkSearch for mIRC
A search window: find networks by keyword or tag, best match first, sort them, see their details and connect with a click. Needs mIRC 7.56 or newer.
Install: save the file, then in mIRC type /load -rs erksearch.mrc and open it with /erksearch.
erkSearch for HexChat
Commands to search networks by keyword or tag, browse the biggest, read their details and connect by number. Needs HexChat's Lua plugin and curl.
Install: put the file in HexChat's addons folder (or /lua load it), then type /erksearch libera.
erkSearch for irssi
The same commands for irssi: search networks by keyword or tag, browse the biggest, read their details and connect by number. Searches run in the background, so irssi never stalls.
Install: save the file to ~/.irssi/scripts/, type /script load erksearch, then /erksearch libera.
Built with the erkstats API
Built a client feature, bot, script or site on the API? Tell us about it and we'll list it here.
Reference
A free, keyless read API for the data you see on erkstats: IRC networks and channels, how to connect to them, and how busy they are. Use it to add a "find a network" feature to an IRC client, show live numbers on your site, or build a bot that reports on your favourite network.
- Base URL:
https://erkstats.info/api/v1/public - Method:
GETonly - Key: none needed
- Format: JSON (
Content-Type: application/json; charset=utf-8) - CORS: open (
Access-Control-Allow-Origin: *), so you can call it from a web page - Caching: responses send
Cache-Control: public, max-age=300(5 minutes) - Rate limit: 120 requests a minute per IP address
Name your app
Please tell us what is calling, with an X-Client header (or a User-Agent):
X-Client: my-irc-client/2.1
It is optional, but it lets us see which apps use the API, keep them working when things change, and get in touch if something goes wrong. If you build something people can use, tell us and we will list it on the API page.
Finding networks and channels
GET /networks
Browse networks: the top N by a sort, or the whole list.
| Param | Type | Default | Notes |
|---|---|---|---|
sort | string | users | One of users, users_30d, channels, servers, founded, name. |
dir | string | desc | asc or desc. (Use sort=name&dir=asc for A to Z.) |
lang | string | - | Only networks that list this language, e.g. lang=german. |
verified | int | - | verified=1 returns only networks that have opted in. |
limit | int | 25 | 1 to 500. Use limit=500 to get them all. |
offset | int | 0 | For paging. |
users is the count right now and moves around through the day; users_30d (the 30-day average) gives a steadier "biggest networks" order.
GET /api/v1/public/networks?sort=users_30d&limit=3
{
"count": 154,
"limit": 3,
"offset": 0,
"sort": "users_30d",
"dir": "desc",
"results": [ /* network objects, see below */ ]
}
GET /search
Search networks and channels together, with the same results as the Search page. It finds:
- Different forms of a word:
gamesalso findsgaming. - Related words:
sciencealso finds biology, physics and mathematics. - Tags and languages: a network's tags and languages find the network and its busiest channels.
- Every word you give: results that match more of them rank higher.
- Channel names: a search like
#linuxputs that exact channel first.
Results come back best match first. Both lists are bigger networks and busier channels first where the match is equally good.
| Param | Type | Default | Notes |
|---|---|---|---|
q | string | - | What to search for, at least 2 characters. Up to 6 words are used. |
tag | string | - | Only networks with this tag, and their channels, e.g. tag=gaming. q can be left out when this is given. |
sort | string | relevance | relevance (best match), users or name. |
dir | string | - | asc or desc. The default is desc, but asc (A to Z) for name. |
network_limit | int | 10 | How many networks to return, 1 to 100. |
limit | int | 25 | How many channels to return, 1 to 100. |
offset | int | 0 | For paging through the channels. |
GET /api/v1/public/search?q=science&network_limit=2&limit=1
{
"query": "science",
"tag": null,
"sort": "relevance",
"dir": "desc",
"also_matching": ["biology", "mathematics", "physics", "scientific"],
"networks": {
"count": 1,
"limit": 2,
"results": [ /* network objects */ ]
},
"channels": {
"count": 92,
"limit": 1,
"offset": 0,
"results": [ /* channel objects, each with its network */ ]
}
}
also_matching lists the other words that found results: other forms of your words, related words, and longer words starting with yours. Show it the way the Search page does ("Also matching: biology, physics...") so people can see why a result turned up.
GET /search/network
Just the networks, with paging. It takes the same q and tag as /search, and the lang and verified filters from /networks.
| Param | Type | Default | Notes |
|---|---|---|---|
q | string | - | At least 2 characters (or give tag). |
tag | string | - | Only networks with this tag. |
sort | string | relevance | relevance, or any /networks sort: users, users_30d, channels, servers, founded, name. |
dir | string | desc | asc or desc (not used with relevance). |
lang | string | - | Only networks that list this language. |
verified | int | - | verified=1 returns only networks that have opted in. |
limit | int | 25 | 1 to 100. |
offset | int | 0 | For paging. |
GET /api/v1/public/search/network?q=libera
Returns a list (see Conventions) with query, tag, sort, dir and also_matching, and network objects as results.
GET /search/channel
Just the channels, with paging, across every network (or one).
| Param | Type | Default | Notes |
|---|---|---|---|
q | string | - | At least 2 characters (or give tag). |
tag | string | - | Only channels on networks with this tag. |
network | string | - | Only channels on this network, by its slug, e.g. network=libera. |
sort | string | relevance | relevance, users or name. |
dir | string | - | asc or desc. The default is desc, but asc for name. |
limit | int | 25 | 1 to 100. |
offset | int | 0 | For paging. |
Each result is a channel object with its network inside it.
GET /api/v1/public/search/channel?q=linux&limit=1
{
"query": "linux",
"tag": null,
"count": 1083,
"limit": 1,
"offset": 0,
"sort": "relevance",
"dir": "desc",
"also_matching": ["debian", "ubuntu", "unix"],
"results": [
{
"name": "#linux",
"topic": "Welcome to #linux! ...",
"users": 2452,
"users_30d_avg": 2495,
"channel_stats": false,
"connect_uri": "ircs://irc.libera.chat:6697/#linux",
"network": { /* network object */ }
}
]
}
Network and channel stats
The numbers behind each network and channel page. <slug> is the network's slug from any network object (it is also the last part of its page address, e.g. erkstats.info/network/libera).
GET /networks/<slug>
One network, with its current stats.
GET /api/v1/public/networks/libera
{
"slug": "libera",
"name": "Libera.Chat",
"server": "irc.libera.chat",
"port": 6697,
"tls": true,
"users": 30429,
"users_30d_avg": 31072,
"connect_uri": "ircs://irc.libera.chat:6697/",
"url": "https://erkstats.info/network/libera",
"stats": {
"users": 30429,
"channels": 22787,
"servers": 30,
"users_change_24h_pct": -0.4,
"users_trend_30d_pct": -3.8,
"rank": 1,
"rank_of": 154,
"peak": { "users": 35273, "at": "2026-08-15T09:25:48Z" },
"tracked_since": "2026-07-25T06:37:13Z",
"updated_at": "2026-11-04T04:08:45Z"
}
}
(Shortened: it carries every field of the network object, plus url and stats.)
| Stat | Meaning |
|---|---|
users | People online at the last check. |
channels | Channels the network reported at the last check. |
servers | Servers in the network. |
users_change_24h_pct | Change in people online against the same time yesterday. |
users_trend_30d_pct | Change over the last 30 days. |
rank, rank_of | Position by people online, out of all the networks we track. |
peak | The most people we have ever seen online, and when. |
tracked_since | Our first reading for this network. |
updated_at | When the numbers were last checked (every few hours). |
GET /networks/<slug>/history
People online over time.
| Param | Type | Default | Notes |
|---|---|---|---|
range | string | 30d | One of 24h, 7d, 30d, 1y, 10y, all. |
from | string | - | Instead of range: a start date (2026-10-01, UTC) or unix time. |
to | string | now | With from: an end date (inclusive) or unix time. |
GET /api/v1/public/networks/libera/history?range=7d
{
"network": "libera",
"range": "7d",
"from": null,
"to": null,
"points": [
{ "t": "2026-10-28T07:13:26Z", "users": 30908 },
{ "t": "2026-10-28T09:14:18Z", "users": 31490 }
]
}
You get up to about 180 points whatever the range: over long ranges, nearby readings are averaged together.
GET /networks/<slug>/channels
The network's busiest channels right now.
| Param | Type | Default | Notes |
|---|---|---|---|
limit | int | 25 | 1 to 100. |
{
"network": "libera",
"listed": true,
"count": 6890,
"limit": 3,
"results": [
{
"name": "#linux",
"topic": "Welcome to #linux! ...",
"users": 2374,
"channel_stats": false,
"connect_uri": "ircs://irc.libera.chat:6697/#linux",
"url": "https://erkstats.info/channel/libera/#linux"
}
]
}
Some networks ask us not to publish their channel list. For those, listed is false and results is empty.
GET /networks/<slug>/channel?name=<channel>
One channel, with its 30-day average, its peak and when we first and last saw it. Put the channel name in name, with the # written as %23:
GET /api/v1/public/networks/libera/channel?name=%23linux
{
"name": "#linux",
"topic": "Welcome to #linux! ...",
"users": 2374,
"channel_stats": false,
"connect_uri": "ircs://irc.libera.chat:6697/#linux",
"url": "https://erkstats.info/channel/libera/#linux",
"users_30d_avg": 2396,
"peak": { "users": 2535, "at": "2026-07-26T19:20:01Z" },
"first_seen": "2026-07-25T06:37:13Z",
"last_seen": "2026-11-04T04:08:45Z"
}
GET /networks/<slug>/channel/history?name=<channel>
People in a channel over time. Takes the same range, from and to as the network history, and returns the same shape with a channel field.
Objects
Network object
| Field | Type | Description |
|---|---|---|
slug | string | Stable identifier. Use it rather than name to link or store a network. |
name | string | Display name. |
server | string | The public server address to connect to. |
port | int | Port. |
tls | bool | Whether the port uses TLS. |
website | string or null | The network's website. |
webchat | string or null | A web page where you can chat without installing a client. |
founded | int or null | Year founded. |
language | string or null | Main language (e.g. "English"), the first of languages. |
languages | string[] | Every language the network lists, main one first (may be empty). |
blurb | string or null | Short description. |
users | int or null | People online at the last check. |
users_30d_avg | int or null | Average people online over the last 30 days. |
verified | bool | The network has confirmed it is happy to be listed (see below). |
socials | object or null | Links to the network's social accounts, e.g. {"mastodon": "https://..."}. |
connect_uri | string | Ready to open: ircs://host:port/ (or irc:// without TLS). |
Channel object
| Field | Type | Description |
|---|---|---|
name | string | Channel name, including the leading #. |
topic | string or null | Latest topic. |
users | int or null | People in the channel at the last check. |
users_30d_avg | int or null | Average over the last 30 days. |
channel_stats | bool | true if erkstats has a full statistics page for this channel. |
connect_uri | string | Ready to open: ircs://host:port/#channel. |
network | object | The network it is on (search results only). |
connect_uri keeps the # of a channel as it is (e.g. ircs://irc.libera.chat:6697/#linux), which is what IRC clients expect. Hand it straight to your client.
verified means the network's operators have confirmed they are happy to be listed and checked by us. Unverified networks are still tracked and returned; the flag just lets you highlight the ones that opted in.
Conventions
Lists come back as:
{ "count": 154, "limit": 25, "offset": 0, "results": [ ... ] }
count is the total number of matches, not the number returned, so you can page with limit and offset until offset + results.length >= count. Searches add "query", "tag" and "also_matching"; the network lists and searches add "sort" and "dir".
Times are UTC, written like 2026-11-04T04:08:45Z.
Errors are {"error": "..."} with an HTTP status:
| Status | When |
|---|---|
400 | A required parameter is missing or wrong (e.g. q too short). |
404 | Unknown network, channel or endpoint. |
405 | A method other than GET. |
429 | Too many requests. Wait for the Retry-After seconds. |
503 | /search only: search is starting up after an update. Try again in a few minutes. |
Rate limit. 120 requests a minute per IP address. Numbers only change every few minutes, so cache what you fetch and you will stay well under it.
Notes
- Freshness. Network numbers are checked every few hours; channel lists about as often. Searches return channels seen in the last week, so closed channels drop out. The word list behind search is rebuilt every hour, so a brand new channel or topic can take up to an hour to be found.
- Matching is by whole words: names, descriptions, tags and languages for networks; names and topics for channels. Case doesn't matter. A word found nowhere is also tried as the start of longer words (
astrofindsastronomy). Searching on tags alone (noq) returns at most 1,000 channels. - Stability. Field names here will not change. New fields may be added over time, so ignore any you don't recognise rather than failing on them.
- Private analytics. Members who manage a network can create a personal API key for their network's private analytics (peak hours, channel churn, daily history). See the member API.