About
Discord CLI for AI agents and humans. Manage servers, send messages, react, handle DMs, threads, and monitor events from the terminal. Built with security profiles, audit logging, and rate limiting for safe AI agent integration.
Documentation
discli
Discord CLI for AI agents and humans
Manage Discord servers, send messages, react, handle DMs, threads, and monitor events, all from the terminal. Built with security and AI agent integration in mind.
How it works
graph TD
A[discli CLI] --> B[Commands]
A --> C[Listen]
A --> D[Security]
B --> |fire & exit| E[discord.py]
C --> |stay connected| E
D --> D1[Permission Profiles]
D --> D2[Audit Logging]
D --> D3[Rate Limiting]
D --> D4[User Permission Check]
E --> F[Discord Bot API]
style A fill:#5865F2,color:#fff
style F fill:#5865F2,color:#fff
style D fill:#ED4245,color:#fff
graph LR
A[discli listen --json] --> B[Events\nJSONL stream]
B --> C[AI Agent\ne.g. Claude]
C --> D[discli message reply]
D --> E[Discord]
A --> E
style A fill:#5865F2,color:#fff
style C fill:#D97706,color:#fff
style E fill:#5865F2,color:#fff
graph TD
subgraph Security Layer
P[Permission Profiles] --> |full / chat / readonly| CHECK{Allowed?}
CHECK --> |Yes| EXEC[Execute Command]
CHECK --> |No| DENY[Deny + Log]
EXEC --> CONFIRM{Destructive?}
CONFIRM --> |Yes| PROMPT[Confirm or --yes]
CONFIRM --> |No| RUN[Run]
PROMPT --> RUN
RUN --> AUDIT[Audit Log]
RUN --> RATE[Rate Limiter]
end
style P fill:#ED4245,color:#fff
style DENY fill:#ED4245,color:#fff
style AUDIT fill:#57F287,color:#000
Install
# macOS/Linux
curl -fsSL https://raw.githubusercontent.com/DevRohit06/discli/main/installers/install.sh | bash
# Windows (PowerShell)
irm https://raw.githubusercontent.com/DevRohit06/discli/main/installers/install.ps1 | iex
Or with pip:
pip install discord-cli-agent
Requires Python 3.10+.
Setup
- Create a bot at Discord Developer Portal
- Enable all privileged intents (Presence, Server Members, Message Content)
- Add the bot to your server with appropriate permissions
- Configure your token:
# Option A: Save to config
discli config set token YOUR_BOT_TOKEN
# Option B: Environment variable
export DISCORD_BOT_TOKEN=YOUR_BOT_TOKEN
# Option C: Pass directly
discli --token YOUR_BOT_TOKEN server list
Usage
Every command supports --json for machine-readable output.
Messages
discli message send #general "Hello world!"
discli message send #general "Check this out" --embed-title "News" --embed-desc "Big update"
discli message list #general --limit 20
discli message list #general --after 2026-03-01 --before 2026-03-14
discli message get #general 123456789
discli message reply #general 123456789 "Thanks for your question!"
discli message edit #general 123456789 "Updated text"
discli message delete #general 123456789
Search & History
# Search messages by content
discli message search #general "bug report" --limit 100
discli message search #general "help" --author alice --after 2026-03-01
# Deep history backfill
discli message history #general --days 7
discli message history #general --hours 24 --limit 500
Direct Messages
discli dm send alice "Hey, need help?"
discli dm send 123456789 "Sent by user ID"
discli dm list alice --limit 10
Reactions
discli reaction add #general 123456789 👍
discli reaction remove #general 123456789 👍
discli reaction list #general 123456789
Channels
discli channel list --server "My Server"
discli channel create "My Server" new-channel --type text
discli channel create "My Server" voice-room --type voice
discli channel info #general
discli channel delete #old-channel
Threads
discli thread create #general 123456789 "Support Ticket"
discli thread list #general
discli thread send 987654321 "Following up on your issue"
Servers
discli server list
discli server info "My Server"
Roles
discli role list "My Server"
discli role create "My Server" Moderator --color ff0000
discli role assign "My Server" alice Moderator
discli role remove "My Server" alice Moderator
discli role delete "My Server" Moderator
Members
discli member list "My Server" --limit 100
discli member info "My Server" alice
discli member kick "My Server" alice --reason "Spam"
discli member ban "My Server" alice --reason "Repeated violations"
discli member unban "My Server" alice
Typing Indicator
discli typing #general # 5 seconds (default)
discli typing #general --duration 10 # 10 seconds
Live Event Monitoring
# Listen to everything
discli listen
# Filter by server/channel
discli listen --server "My Server" --channel #general
# Filter by event type
discli listen --events messages,reactions
# Include bot messages (ignored by default)
discli listen --include-bots
# JSON output for piping to an agent
discli --json listen --events messages
Supported event types: messages, reactions, members, edits, deletes
Security & Permissions
Confirmation Prompts
Destructive actions (kick, ban, delete) require confirmation by default:
$ discli member kick "My Server" spammer
Warning: Destructive action: member kick (spammer from My Server). Continue? [y/N]
# Skip with --yes for automation
$ discli -y member kick "My Server" spammer --reason "Spam"
Permission Profiles
Restrict which commands an agent can use:
# List available profiles
discli permission profiles
# Set a profile
discli permission set chat # Messages, reactions, threads only, no moderation
discli permission set readonly # Can only read, no sending or deleting
discli permission set moderation # Full access including kick/ban
discli permission set full # Everything (default)
| Profile | Can Send | Can Delete | Can Kick/Ban | Can Manage Channels |
|---|---|---|---|---|
full | Yes | Yes | Yes | Yes |
moderation | Yes | Yes | Yes | Yes |
chat | Yes | No | No | No |
readonly | No | No | No | No |
User Permission Checking
Verify the Discord user who triggered an action actually has the required permissions:
discli member kick "My Server" target --triggered-by 123456789
This checks that user 123456789 has kick_members permission in the server. Server owners and administrators always pass. If the user can't be found in cache, it fetches from the API. If that also fails, it warns but doesn't block.
Audit Log
Every destructive action is logged to ~/.discli/audit.log:
# View recent actions
discli audit show --limit 20
# JSON output
discli --json audit show
# Clear the log
discli audit clear
Rate Limiting
Built-in rate limiter (5 calls per 5 seconds) on destructive actions to prevent Discord API bans. If the limit is hit, discli waits automatically.
Resolving Identifiers
All commands accept both IDs and names:
| Type | By ID | By Name |
|---|---|---|
| Channel | 123456789 | #general |
| Server | 123456789 | My Server |
| Member | 123456789 | alice |
| Role | 123456789 | Moderator |
| Thread | 123456789 | Support Ticket |
| User (DM) | 123456789 | alice |
JSON Output
Add --json before the subcommand for machine-readable output:
$ discli --json message list #general --limit 1
[
{
"id": "123456789",
"author": "alice",
"content": "Hello!",
"timestamp": "2026-03-14T10:32:00+00:00",
"attachments": [],
"embeds": []
}
]
$ discli --json listen --events messages
{"event": "message", "server": "My Server", "channel": "general", "channel_id": "111", "author": "alice", "author_id": "222", "content": "hello", "message_id": "333", "mentions_bot": false, "attachments": []}
Examples
Ready-to-run examples in the examples/ directory:
| Example | Description |
|---|---|
claude_agent.py | AI support agent powered by Claude Agent SDK with persistent session |
support_agent.py | Keyword-based support bot that replies to @mentions |
thread_support_agent.py | Creates a thread per support request and continues conversations inside |
moderation_bot.py | Watches for banned words, warns users, kicks after repeated violations |
channel_logger.sh | Logs all messages from a channel to a JSONL file |
reaction_poll.sh | Creates a poll with emoji reactions |
Agent Instructions
The agents/discord-agent.md file contains the full discli command reference for AI agents. Drop it into any agent's system prompt. Works with Claude, OpenAI, LangChain, or any framework.
Quick start: Claude Agent
pip install discord-cli-agent claude-agent-sdk
discli config set token YOUR_BOT_TOKEN
python examples/claude_agent.py
Uses your existing Claude Code authentication. No API key needed.
Quick start: Bash agent loop
discli --json listen --events messages | while read -r event; do
mentions_bot=$(echo "$event" | jq -r '.mentions_bot')
if [ "$mentions_bot" = "true" ]; then
channel_id=$(echo "$event" | jq -r '.channel_id')
message_id=$(echo "$event" | jq -r '.message_id')
discli typing "$channel_id" --duration 3 &
discli message reply "$channel_id" "$message_id" "Hello! How can I help?"
fi
done
Project Structure
discli/
├── src/discli/
│ ├── cli.py # Root click group + permission/audit commands
│ ├── client.py # Async discord.py wrapper
│ ├── config.py # Token storage (~/.discli/config.json)
│ ├── security.py # Permissions, audit logging, rate limiting
│ ├── utils.py # Output formatting, resolvers
│ └── commands/ # Command groups (message, channel, role, etc.)
├── agents/
│ └── discord-agent.md # Full command reference for AI agents
├── examples/ # Ready-to-run agent examples
├── installers/ # curl-friendly install scripts
├── tests/ # Unit tests
└── pyproject.toml
Configuration
Config is stored at ~/.discli/config.json.
discli config set token YOUR_TOKEN
discli config show
discli config show --json
Token resolution order: --token flag > DISCORD_BOT_TOKEN env var > config file.
Contributing
See CONTRIBUTING.md for development setup, how to add commands, and release process.
License
MIT