An open API service indexing awesome lists of open source software.

https://github.com/zebbern/claude-code-discord

A Discord bot that brings Claude Code to your channels so you can chat, run shell/git, and manage branches. Access from any local, VM, or Docker instance with full capebilty of claude code!
https://github.com/zebbern/claude-code-discord

List: claude-code-discord

anthropic anthropic-claude automation-framework awesome awesome-list awesome-lists claude claude-ai claude-ai-bot claude-code claude-code-commands claude-code-discord developer discord-api discord-bot opus sonnet-4 tools-and-automation

Last synced: 6 months ago
JSON representation

A Discord bot that brings Claude Code to your channels so you can chat, run shell/git, and manage branches. Access from any local, VM, or Docker instance with full capebilty of claude code!

Awesome Lists containing this project

README

          

# claude-code-discord

| Advantage | Details | Status |
|---------------------------------------------|---------------------------------------------------------------------------------------------|:------:|
| Use Claude Code Anywhere | Host locally (VM / Docker / cloud) and send commands via the Discord API | ✅ |
| Centralized collaboration | Run commands and discuss results where your team already communicates | ✅ |
| Branch-aware organization | Maps Git branches to channels/categories so feature work stays separated | ✅ |
| Immediate, shareable feedback | Execute `/git`, `/shell`, or `/claude` and get outputs directly in-channel | ✅ |
| Reduced context switching | Keep actions, logs, and discussion together — less switching between terminal and chat | ✅ |
| Role-based access control | Restrict destructive commands (`/shell`, `/git`, worktree ops) to specific Discord roles | ✅ |
| Non-developer friendly | PMs, QA, and stakeholders can trigger checks and view results without repo access | ✅ |
| Automatable touchpoint | Integrate with CI/webhooks to triage issues and run fixes from chat | ✅ |
| Local hosting & security | Keep keys and code on your infra while exposing a controlled interface through Discord | ✅ |
| Audit trail & accountability | Channel history provides an easy-to-search record of who ran what and when | ✅ |


**Start Here If You Have These:**
DISCORD_TOKEN
APPLICATION_ID

- **[Quick Start](#pre)**
- **[Command List 48 Commands](#Command-list)**

**Tutorial If you dont know or have them follow these first then come back:**
- **[How To Setup Discord Bot?](#setup)**

---

### Preview:
image

Quick Start

- **Install Deno Can Be Done Via [Denos Website](https://deno.com/) Or Commands Under:**
```C
# Linux/MacOS
curl -fsSL https://deno.land/install.sh | sh

# Windows|Powershell
irm https://deno.land/install.ps1 | iex
```

**Clone the project:**
```
git clone https://github.com/zebbern/claude-code-discord.git
cd claude-code-discord
```
**Install claude `If you dont have it` and login:**
```
npm install -g @anthropic-ai/claude-code
claude /login
```
**Required environment variables**
```
# Linux
export DISCORD_TOKEN="your-discord-bot-token"
export APPLICATION_ID="your-discord-app-id"

# Windows Terminal
set DISCORD_TOKEN=your-discord-bot-token
set APPLICATION_ID=your-discord-app-id

# Windows Powershell
$env:DISCORD_TOKEN = "your-discord-bot-token"
$env:APPLICATION_ID = "your-discord-app-id"
```
**Now Run the discord Bot**
> If you get `not a git directory` just run
> git add .
```bash
# Option 1: Just run the bot
deno run --allow-all index.ts

# Option 2: Run the bot with extra terminal output (For Dev)
deno run --allow-all index.ts --watch

# Option 3: Run the bot and get pings when claude is done
./index.ts --category myproject --user-id Your_Discord_User_ID_Here

# Can also be ran like this:
deno run --allow-all index.ts --category yourproject --user-id Your_Discord_User_ID_Here
```
**You can run without `--user-id Your_Discord_User_ID_Here` if you dont want to be notified when claude finishes**
image

Setup Discord Bot

1. Create a Discord Application

> [!Note]
> - Go to the [Discord Developer Portal](https://discord.com/developers/applications)
> - Click New Application
> - Give your application a name (e.g., ClaudeCode)
> - Click Create
> app-create

2. Copy Application ID (Needed For Config)

> [!Note]
> - Go to the General Information → Copy Application ID section
> APPLICATION_ID

3. Create a Bot User

> [!Note]
> - In your application, go to the Bot section in the left sidebar
> - Click Add Bot
> - Under Token click Copy to copy your bot token (keep this secure!)
> - Click Save Changes
> image

4. Invite the Bot to Your Server

> [!Note]
> - Go to the OAuth2URL Generator section
> - Under Scopes select:
> ```
> + | bot
> + | applications.commands
> ```
> - Under Bot Permissions select:
> ```
> + | Send Messages
> + | Use Slash Commands
> + | Read Message History
> + | Embed Links
> ```
> Copy the generated URL and open it in your browser
> Select your Discord server and authorize the bot
> oauth2
> botallowcommands
> image

## Command List
> (48 Commands)

### Core Claude (3)
- `/claude`, `/continue`, `/claude-cancel`

### Enhanced Claude (4)
- `/claude-enhanced`, `/claude-models`, `/claude-sessions`, `/claude-context`

### Development Tools (7)
- `/claude-explain`, `/claude-debug`, `/claude-optimize`, `/claude-review`
- `/claude-generate`, `/claude-refactor`, `/claude-learn`

### New Features (3)
- `/todos` - Task management with API rate limits
- `/mcp` - Model Context Protocol servers
- `/agent` - 7 specialized AI agents

### Settings (4)
- `/settings` - Unified settings (NEW)
- `/claude-settings`, `/output-settings`, `/quick-model`

### Git Operations (6)
- `/git`, `/worktree`, `/worktree-list`, `/worktree-remove`
- `/worktree-bots`, `/worktree-kill`

### Shell Management (4)
- `/shell`, `/shell-input`, `/shell-list`, `/shell-kill`

### System Monitoring (10)
- `/system-info`, `/processes`, `/system-resources`, `/network-info`
- `/disk-usage`, `/env-vars`, `/system-logs`, `/port-scan`
- `/service-status`, `/uptime`

### Utilities (4)
- `/status`, `/pwd`, `/shutdown`, `/help`

### Agent System (3)
- `/agent` with 7 specialized agents:
- Code Reviewer, Software Architect, Debug Specialist
- Security Analyst, Performance Engineer, DevOps Engineer, General Assistant

#### Thinking Mode Options ✨
- `none` - Standard Claude responses
- `think` - Step-by-step reasoning mode
- `think-hard` - Deep analysis and reasoning
- `ultrathink` - Maximum depth thinking for complex problems

#### Operation Mode Options ✨
- `normal` - Standard operation with user confirmation
- `plan` - Planning mode without execution
- `auto-accept` - Automatically apply suggested changes
- `danger` - Unrestricted mode (high risk)

#### `/todos` Command ✨
- **Action types**: list, add, complete, generate, prioritize, rate-status
- **Priority levels**: low, medium, high, critical
- **Rate limit awareness** - Supports Anthropic API tiers including `exceeds_200k_tokens`
- **Token estimation** - Calculates estimated token usage
- **Auto-generation** - Generate todos from code files

#### `/mcp` Command ✨
- **MCP server management** - Model Context Protocol integration
- **Server types**: local, http, websocket, ssh
- **Actions**: list, add, remove, test, status
- **Connection testing** - Verify MCP server connectivity

#### `/agent` Command ✨
- **Specialized AI agents** for different development tasks:
- Code Reviewer - Quality analysis and security
- Software Architect - System design and architecture
- Debug Specialist - Bug analysis and troubleshooting
- Security Analyst - Vulnerability assessment
- Performance Engineer - Optimization and profiling
- DevOps Engineer - Deployment and infrastructure
- General Assistant - Multi-purpose development help
- **Risk levels** - Low/Medium/High risk classification
- **Session management** - Persistent agent conversations
- **Context awareness** - Include system info and files