Discord bot market statistics

Discord’s own App Directory lists 10,121 apps under Moderation and Tools, MEE6 ranks 21st in it, and a single-purpose colour-role bot outranks MEE6. Entry paid prices cluster between $2.99 and $7.99 across nine different billing units, and a quarter of the products examined publish no price at all.

What this page is

Every figure below comes from a study of 104 Discord bot products examined in depth, plus roughly 190 more catalogued, carried out on and around 30 August 2026. Each number names its source.

Where a figure is Discord’s own, it is marked as such. Where it is a vendor’s published price, it is a list price on the date checked, not a promotional one. Where something could not be determined, that is stated rather than estimated.

This is a sample, not a census. Discord’s directory lists more than ten thousand apps in one category alone.

How big the market actually is

Discord’s App Directory publishes its own category counts through /api/v9/application-directory-static/categories, which answers an unauthenticated request. Read 30 August 2026:

Everything in the directory is Discord-verified by definition, which makes these counts and the ordering the closest thing to an authoritative index that exists — it is Discord’s own, not a vendor’s self-reported figure.

One structural fact worth knowing before you look. The public page at discord.com/discovery/applications is now entirely games and Activities — promoted games, puzzle games, card games, watch-together, social activities, and a four-item “Bot Games” row. A logged-out visitor is shown no moderation, utility or server-management apps at all. The catalogue is still public, but it lives behind the JSON API rather than on that page.

What Discord’s own ranking shows

The top 25 in Moderation and Tools, in Discord’s order, read 30 August 2026:

FlaviBot · TTS Bot · carl-bot · Arcane · Dyno · Bloxlink · DISBOARD.org · Mimu · Maki.gg · Make it a Quote · Ticket Tool · Invite Tracker · Koya · ProBot · Loritta · Appy · Confessions · Phoenix · Color-Chan · DraftBot · MEE6 · Tickets v2 · Security · Giveaway Boat · Statbot

Four things in that list contradict the standard English listicle.

MEE6 ranks 21st. The bot every “best Discord bots” article leads with is not what Discord itself surfaces first.

Color-Chan outranks MEE6. A single-purpose colour-role bot sits at 19.

A text-to-speech bot ranks 2nd, and a server-discovery listing ranks 7th — neither of which is a moderation bot in any ordinary sense.

DraftBot, a French all-in-one with its own dashboard and full documentation, ranks 20th — above MEE6 — and is absent from essentially every English-language source. So is RaidProtect. The French-speaking market has two products at scale that English coverage does not mention.

Ranks 36 and 38 are link-filter bypass tools, verified and listed inside Discord’s own moderation category. Any anti-scam claim has to survive tools that are themselves directory-listed.

Installed-server figures

These are the directory’s own reported figures, read from Discord’s API on 30 August 2026. They are what Discord publishes, not an independent audit, and one vendor’s site disagrees with the directory by about 6,000.

The point of that table is not any single row. It is that single-purpose bots reach install numbers that the all-in-one narrative does not account for. A counting bot is on 1.28 million servers.

Pricing

The range

Entry paid tiers across the products examined run from $1 per month to $149 per month, and the top and bottom are both instructive.

Prices cluster hard between $2.99 and $7.99. Everything above $12 is either an analytics product, a security product priced by member count, or a platform tier.

Nine billing units, and why headline prices do not compare

A $5 subscription covering unlimited servers and a $5 subscription covering one are not the same product, and a comparison table that renders them as the same number is wrong.

Currencies observed: USD, GBP, EUR, BRL and SAR. Any total mixing them is a mixed total; none is converted here.

What a specialist stack costs

Six categories, using the best-regarded specialist in each rather than a strawman. List prices, 30 August 2026:

Substituting the cheapest credible option in each category — Security Bot, Lurkr, Tickets, Giveaway Boat, Streamcord and Statbot’s Top Picks bundle — still costs $26.44 a month, $317.28 a year.

Extending to twelve bots by adding backup, temp voice, applications, reaction roles, reminders, events and alt detection reaches $57.91 plus £8.98 per month, plus one product whose price exists only inside the Discord client.

That mixed-currency total is itself a finding. Two of the twelve bill in GBP, so there is no single correct dollar figure for a twelve-bot stack. A server owner assembling specialists cannot state what they are paying without doing a currency conversion.

The caveat that must travel with the arithmetic: the specialist stack buys depth. Ticket Tool at $7.99 has sixteen per-panel configuration surfaces. Statbot at $14.95 has member and channel drilldowns. The comparison is about cost and coordination, not about specialists being worse at their one job.

A quarter of the market publishes no price

Twenty-six of the products examined cannot be budgeted for from public information. That is roughly a quarter of everything profiled, and the reasons vary:

Four things a buyer therefore cannot find out:

1. Whether the upgrade is worth it. Zira publishes “bigger limits” with no numbers. Giveaway Boat renders an identical benefit list at $3 and at $5.
2. What a unit of AI buys. Vendors publish an allowance and refuse to publish what one unit costs. One vendor’s own pricing page contradicts its own FAQ by a factor of ten on exactly that question.
3. What happens on renewal. MEE6 and Koya both display a first-period price roughly half the renewal.
4. What stops working without votes. Five vendors vote-lock part of the free tier, and not one publishes which commands are affected.

Promotions running on 30 August 2026

Two details worth recording as method rather than as news. Koya’s sale counted down from “1 days, 8 hours” to zero within the same browsing session. And ten vendors advertise annual discounts of 16–40% as standing structure rather than as a promotion — reporting those as sales would be wrong.

Thirty named vendors showed no promotion at all, including Dyno, ProBot, Carl-bot, Streamcord, Zira, Wick, Beemo, Ticket Tool, VoiceMaster, Apollo, Bloxlink, RoWifi and MonitoRSS.

Where the free and paid line falls, by capability

All figures below are counts of products examined in this study, with the sample size named.

The single clearest pattern: enforcement is given away, visibility is sold. Four of sixteen moderation bots paywall logging. Nobody paywalls the ban command.

Three free tiers are shaped in ways a reader will not expect:

1. Security Bot gives away anti-nuke and anti-raid and charges for chat automod — the inverse of everyone else. A free Security Bot server is protected from its own administrators but not from spam.
2. BuildMyDiscord’s free 600 credits are one-time, not monthly — a trial wearing a free tier’s clothes.
3. SCNX’s free tier requires an advertising check-in every 7 days or the plan lapses.

And one shaped right, which deserves saying: Countr gates exactly two commands out of its entire codebase, raises every ceiling on payment, removes nothing from the free tier, and vote-locks nothing.

Discord’s own platform limits

From Discord’s developer documentation, read 30 August 2026. These are platform facts, not a sample.

The six-keyword-rule cap is the single most consequential number in the platform, and it is the reason the third-party automod market exists. Full treatment in AutoMod versus third-party bots.

Developer APIs

The count of APIs rose on every research pass — from none, to three, to six, to nine, to twelve, to thirteen. That is a lesson about absence claims rather than about the market. Detail in which Discord bots publish a developer API.

Localisation

Cakey Bot’s figure is the most rigorously checked in this study, because its translations are in a public repository: an English source of 1,551 keys, every one of the 16 target locales holding 1,538 — a uniform 13-key lag rather than a ragged one — with 96.0% to 99.8% of values differing from English, and sampled Arabic and Turkish reading as idiomatic human translation with interpolation placeholders preserved.

One claim this data does not support is any Western product winning on language against a native-language incumbent. Zaman in Saudi Arabia, GBBOT in Brazil, ErensiBOT in Turkey, DraftBot and RaidProtect in France, Nekotina across Spanish and Portuguese — these products are the language rather than translated into it, and several take local payment rails.

Product mortality

The category has a real attrition rate, and the standard listicle has not caught up.

The nuance in the AltDentifier row generalises: a dead site does not mean a dead listing. Checking one and not the other produces the wrong answer in both directions.

Vetox, as a reference row

Included because a comparison needs a fixed point with published numbers.

On breadth, this is the most complete product in the study, and the closest competitor found is Cakey Bot at 34 modules and 368 commands. The rest of the named field sits considerably lower. Breadth is not the only measure that matters — the capability tables above show several products leading their category outright — but it is the one this study could count directly.

Method and scope

Sample sizes by domain, which every absence statement above is scoped to:

Three method notes worth reusing.

Enumerate a documentation site from its sitemap.xml, never from its navigation. One vendor’s audit-log feature was recorded as absent because the rendered nav omits the section it lives in; the sitemap listed three unrecorded feature pages.

Discord’s App Directory API is the cheapest falsification tool available. It answers unauthenticated requests, returns exact JSON, and settles listing, verification status and rank without a browser.

Every claim this study had to withdraw was an absence claim, and every one failed the same way: the scope was wider than the evidence. “No Discord bot does X” is falsifiable by anyone with five minutes and a directory search. “Of the sixteen moderation bots we checked, two do X” is not.