https://github.com/makesnosense/voice
A minimalistic, self-hostable WebRTC voice chat.
https://github.com/makesnosense/voice
self-hosted socket-io voice-chat webrtc
Last synced: 12 days ago
JSON representation
A minimalistic, self-hostable WebRTC voice chat.
- Host: GitHub
- URL: https://github.com/makesnosense/voice
- Owner: makesnosense
- License: mit
- Created: 2025-09-05T20:21:52.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2026-01-10T12:50:28.000Z (14 days ago)
- Last Synced: 2026-01-11T04:00:01.063Z (14 days ago)
- Topics: self-hosted, socket-io, voice-chat, webrtc
- Language: TypeScript
- Homepage:
- Size: 531 KB
- Stars: 5
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
**A minimalistic, self-hostable WebRTC voice chat.**
## Built with
**Client:** React 19, TypeScript, Vite, Zustand, Socket.IO Client
**Server:** Express, Socket.IO Server, Bun
**Infra:** Docker, nginx, coturn, Let's Encrypt
## Self-hosting
### Prerequisites
- **Server**: Linux VPS with 1 GB RAM or more
- **Domain**: Any domain pointing to your server's IP (free options: [FreeDNS](https://freedns.afraid.org), [DuckDNS](https://www.duckdns.org))
- **Docker**
- **Open Ports**: 80 (HTTP), 443 (HTTPS), 3478 (TURN server), 49152-49172 (TURN server UDP range)
### Installation
#### 1. Clone the repository
```bash
git clone --branch v1.0-selfhost --depth 1 https://github.com/makesnosense/voice.git
cd voice
```
This downloads the stable self-hosted version (`v1.0-selfhost`) without full git history.
#### 2. Set up environment
```bash
cp .env.selfhost.example .env
nano .env # or use vim, micro, etc.
```
Edit the following in `.env`:
```bash
DOMAIN=your-domain.com # your actual domain
EMAIL=your-email@example.com # for Let's Encrypt notifications
COTURN_SECRET= # generate with: openssl rand -base64 32
```
To generate a strong secret:
```bash
openssl rand -base64 32
```
#### 3. Choose setup method
##### Option A: Quick Setup
If you don't mind using a setup script, this is the fastest way. (setup.sh is just two `docker run` commands)
```bash
bash setup.sh
docker compose up -d
```
That's it! ✨
Your voice chat is now running at `https://your-domain.com`
##### Option B: Manual Setup
###### 1. Get SSL certificate
Load environment variables and get certificate:
```bash
source .env
docker run --rm -p 80:80 \
-v voice_certbot-conf:/etc/letsencrypt \
certbot/certbot certonly --standalone \
--email ${EMAIL} \
--agree-tos --no-eff-email --non-interactive \
-d ${DOMAIN}
```
You should see: `Successfully received certificate`
###### 2. Build frontend
```bash
source .env
docker run --rm \
-v ./client:/app/client \
-v ./shared:/app/shared \
-v voice_client-dist:/app/dist \
-e VITE_TURN_SERVER_HOST=${DOMAIN} \
-e VITE_TURN_SERVER_PORT=${VITE_TURN_SERVER_PORT} \
-w /app/client \
node:alpine \
sh -c "npm ci && npm run build && cp -r dist/* /app/dist/"
```
This takes 1-2 minutes. You should see: `✓ built in XXs`
###### 3. Start services
```bash
docker compose up -d
```