Kizuna

Self-hosted Discord alternative with text chat, voice channels, and screen sharing.
You host the server, you own the data.

Important

Voice channels rely on WebRTC, which may have limited support on mobile browsers and some Linux configurations. If you run into issues, please open a GitHub issue or reach out — your feedback helps improve cross-platform support.

Features

• Real-time text chat — channels, direct messages, typing indicators, and @mentions (@everyone, @here, @user)
• Voice channels — powered by WebRTC via mediasoup SFU with per-channel audio quality controls
• Screen sharing — share your screen in voice channels (desktop client only)
• Custom roles & permissions — create roles with granular permissions (send messages, manage channels, delete messages, kick members, manage invites)
• File uploads & attachments — images, video, audio, PDFs, and more with drag-and-drop
• Invite codes — generate join links with usage limits and expiry, with QR codes
• Cross-server — connect to multiple self-hosted Kizuna servers from one client
• Desktop client — Windows, macOS (Apple Silicon), and Linux via Tauri v2 with auto-updates and background notifications
• Docker deployment — one-command deploy with automatic HTTPS via Caddy

Project Structure

Package	Path	Purpose
@kizuna/server	apps/server/	Node.js backend, WebSocket API, mediasoup SFU for WebRTC
@kizuna/desktop	apps/desktop/	Tauri v2 desktop client (SvelteKit + Vite)
@kizuna/shared	packages/shared/	Shared TypeScript types and utilities

Deploy

Docker (recommended)

git clone https://github.com/ItsAshn/kizuna.git && cd kizuna
cp .env.example .env
# Edit .env — set DOMAIN and JWT_SECRET at minimum
mkdir -p data uploads && chown -R 1000:1000 data uploads
docker compose up -d

The server will be available at https://your-domain.com. Pre-built images are pulled from ghcr.io/itsashn/kizuna — no compilation or build tools needed on the server. Multi-arch images support amd64 and arm64 (Raspberry Pi, AWS Graviton).

For voice and screen sharing, ensure UDP ports 40000-40099 are reachable on your firewall.

Updating

docker compose pull && docker compose up -d

The compose file includes Watchtower which automatically pulls and applies updates within an hour of a new release. Remove the watchtower service from docker-compose.yml if you prefer manual updates.

For safe manual updates with health-check verification and automatic rollback:

./scripts/update.sh                 # latest version
./scripts/update.sh --pin v0.2.0   # specific version

Local development builds

docker compose -f docker-compose.yml -f docker-compose.dev.yml up -d --build

Manual

# Requires Node.js 22+ and pnpm 9+
pnpm install
pnpm build:server
pnpm start

Environment Variables

All settings are in the .env file. See .env.example for a full reference.

Required:

• DOMAIN — your domain name (Caddy needs this for HTTPS)
• JWT_SECRET — generate with openssl rand -hex 64

Optional but recommended:

• SERVER_NAME — display name shown to clients
• SERVER_DESCRIPTION — short description
• SERVER_URL — full HTTPS URL (used for invite codes)
• SERVER_PASSWORD — optional join password
• PUBLIC_ADDRESS — your server's public IP for WebRTC (auto-detected if blank)

Desktop Client

Download the latest release from the releases page and point it at your server's address.

Development

Quick Start (recommended)

curl -fsSL https://raw.githubusercontent.com/ItsAshn/kizuna/main/scripts/install.sh | bash

This installs Rust, Node.js, pnpm, and all system dependencies (including rebuilding webkit2gtk with WebRTC on Arch). On Windows, it sets up WebView2, Visual C++ Build Tools, and the full toolchain.

Skip the WebRTC rebuild if you only need text chat:

curl -fsSL https://raw.githubusercontent.com/ItsAshn/kizuna/main/scripts/install.sh | bash -s -- --skip-webrtc

Manual Setup

pnpm install

# Start the server (http://localhost:5000)
pnpm dev:server

# Start the desktop client dev server (http://localhost:1420)
pnpm dev:desktop

Open http://localhost:1420 in Chrome or Firefox. Voice channels require WebRTC support.

Linux Voice / WebRTC

Voice channels need webkit2gtk-4.1 compiled with ENABLE_WEB_RTC=ON. Most distros ship without it.

Distro	Status	Fix
Arch / CachyOS (extra)	No WebRTC	scripts/build-webkit-webrtc.sh (rebuilds from PKGBUILD, ~45 min)
Debian / Ubuntu	No WebRTC	scripts/build-webkit-webrtc.sh (rebuilds from apt source, ~60 min)
Fedora	No WebRTC	scripts/build-webkit-webrtc.sh (rebuilds from dnf source, ~60 min)
CI AppImage	WebRTC enabled	Pre-built AppImage bundles patched webkit — voice out of the box

The install script detects your distro and offers to run build-webkit-webrtc.sh automatically. If you skip the rebuild, use pnpm dev:desktop and open Chrome/Firefox instead — voice will work there.

PipeWire systems (Arch, CachyOS, Fedora, etc.): Ensure pipewire-pulse is installed.

sudo pacman -S pipewire-pulse
systemctl --user enable --now pipewire-pulse pipewire-pulse.socket

Build from Source

# One-line setup (recommended)
curl -fsSL https://raw.githubusercontent.com/ItsAshn/kizuna/main/scripts/install.sh | bash

# Or manually — requires Node.js 22+, pnpm 9+, Rust toolchain
pnpm install

# Build the server
pnpm build:server

# Build the desktop client
pnpm build:desktop

The CI-built AppImage and .deb bundle the patched webkit2gtk automatically — voice works out of the box for end users.

License

GNU Affero General Public License v3.0 (AGPLv3). You may use, modify, and distribute this software freely, including for commercial purposes. If you modify the software and make it available as a network service, you must provide the complete source code of your modified version under the same license.