Say it in plain words
Describe the server you want. The agent turns it into real channel, category, role and permission changes, then applies them.
Discord administration, operated by an agent
Tell your server
what to become.
Modu reads the live server, plans each change and asks before anything destructive. Putting the bot in one server and bumping it on Discover is free. A plan turns on the agent and the rest of the bot.
ExampleDeveloper Hub
connectedCreate a professional developer community.
Thinking…
I'll add Development, Community and Staff sections, create a Developer role, and lock staff channels to moderators.
Inspect Discord server420ms
4 categories, 11 channels, 6 roles
Create category310ms
Development
Create role340ms
Developer
Set channel permission overwriteRunning
#staff-internal, deny @everyone
- Typed operations
- 146
- Bot permissions
- 9
- Admin roles granted
- 0
From sign in to a changed server
01
Sign in with Discord
Free. We read your identity and the servers where you hold Manage Server or Administrator.
02
Add the bot
One click puts Modu in one server. From that moment .bump works and your server can sit on Discover, with no card.
03
Start a plan when you want more
The agent, the commands, moderation, tickets and the rest of the bot run on a plan. You see every plan of changes before it runs.
Capabilities
Built to actually operate a server
Not a chatbot that answers questions about Discord. An agent with a fixed set of real operations and a policy layer around them.
Audits that catch what you stopped seeing
A health score with the fix attached: staff channels @everyone can read, roles quietly holding Administrator, duplicates, dormant channels.
A full bot, not just a chat
Moderation, automod, tickets, giveaways, leveling, welcome messages, starboard and join to create voice, all set up by the same agent.
Undo with full history
Every change records what it replaced. Reversible work rolls back in one step, and anything Discord cannot restore is named as such.
82/ 100 server healthExample
- 2 staff channels visible to @everyone
- 6 channels with no activity in 45 days
- 3 duplicate roles
- 5 structural recommendations
Things people ask it to do
- “Set up a professional developer community with rules, roles and a staff area”
- “Audit my server and fix anything @everyone should not see”
- “Create a ticket system for support with a private staff channel”
- “Add a button role menu for pronouns and colour roles”
- “Turn on leveling with rank cards and give Active at level 10”
- “Add 10 aesthetic emojis that fit a cozy anime server”
- “Schedule a weekly announcement every Friday at 6pm”
- “Set up automod against spam, invite links and bad words”
- “Make a join-to-create voice channel”
- “Restyle every bot embed in our brand colour”
Free with the bot
Get found on Discover
Put your server on the public directory with a description, category and tags. Anyone in the server types .bump every 2 hours to move it back to the top. No card, no plan.
- Your own page, written for search engines
- Members bump it, so activity decides the order
- The only thing the bot does for free, and the only thing it needs
Discover
Live now, most recently bumped first
Nothing destructive happens without you
Every operation carries a risk level. Reads and additions run straight away, renames and permission changes follow your setting, and deletions always stop and wait.
- SafeInspect, create channels and categories, add ordinary roles.
- ModerateRename, restructure, change permissions, post messages.
- DangerousDelete channels or roles, grant privileged permissions.
Execution plan
Example1 destructive3 operations
- 01Rename channel to #web-devmoderate
- 02Set channel permissions on #announcementsmoderate
- 03Delete channel #old-generaldangerous
Enforced in code, not asked of the model
Typed operations only
Every action is a named operation with a fixed schema. The model has no raw API access to misuse.
Approvals bound to the plan
An approval covers one execution and a hash of its exact operations, so it cannot be replayed or stretched.
Discord content is data
Channel names, topics and messages are never read as instructions. A prompt hidden in a channel name goes nowhere.
Your account picks the server
The server is resolved from your connection before any tool runs. The model never chooses where it acts.
Powerful bot, strict tools
The bot installs with Administrator so features work first time, but its tools never create or grant Administrator, and risky changes wait for your approval.
Encrypted and audited
OAuth tokens are encrypted with AES-256-GCM at rest. Every change logs its actor, arguments, previous state and result.
What it costs
Free to get listed. Paid to run the server.
Free covers one server, a Discover listing and .bump. Plans start at $9 a month and turn on the agent, the commands, moderation and every other feature.
Questions worth asking
If you are handing an agent write access to a community, you should interrogate it first.
Is there an AI that can manage my Discord server?
Yes. Modu is an AI agent for Discord: you describe what you want in a chat, and it inspects the server, plans the changes and applies them through Discord's API, asking first before anything destructive.
Can AI set up a whole Discord server for me?
Yes. Describe the community (gaming, study, support, creator and so on) or pick one of the templates, and the agent creates the categories, channels, roles and permissions, then configures bot features like welcome messages, tickets and leveling.
Does it analyze my Discord server?
Yes. It audits the live server and returns a health score with fixes: staff channels visible to everyone, roles holding Administrator, duplicate roles, dormant channels and missing moderation.
Is it also a moderation bot?
Yes. The built-in bot runs moderation commands, automod, anti-spam and anti-raid, a jail, an event log and case history, alongside tickets, giveaways, leveling, starboard and VoiceMaster.
What can I do without paying?
Sign in, put the bot in one server, list that server on Discover and bump it every 2 hours. That is the whole free tier: .bump is the only command it answers, and every other feature, including the agent, needs a plan.
Why does the bot ask for Administrator?
So moderation, voice and other features work without re-authorising each time one is added. You can untick it on Discord's screen: the bot then runs on View Channels, Manage Channels, Manage Roles, Send Messages, Manage Messages, Read Message History, Manage Webhooks, Manage Expressions, and Add Reactions, and some features ask for more. Either way its tools refuse to create or grant Administrator roles, and it can only manage roles below its own.
Can I undo what it did?
Reversible work can be rolled back. Created channels and roles, renames, permission overwrites and posted messages record their previous state. Deletions are permanent in Discord, and the product says so.
What are AI credits?
Each plan comes with a monthly allowance of AI credits. A request to the agent uses credits for the AI work it takes: a typical request uses about 18, a quick question less, a whole-server setup more. The bot's own features, like moderation and tickets, never use credits.
Which model runs my requests?
Each plan maps to a model tier. Model access never changes the authorisation rules, which live in backend policy code.
Put it in your server
Adding the bot takes one click and costs nothing. When you want it to rebuild the place instead of just bumping it, the plans are on the pricing page.