Forward user voice from Discord to be accessible on web

Limitation/Known Issue ⚠

• Memory level stayed high when stress test stopped (possible leak). Refer to server-side test result

Important ❗

Please read notes/README.txt

Features 📜

1. Capture whitelisted users or entire voice channel (anyone) voice
2. Realtime Audio Processing using FFMPEG
3. Raw/Processed audio stream exposed on endpoints
4. HTTP, HTTPS server using Express
5. WebSocket using Socket.IO
6. WebRTC transmission

Tests 🔧

Found at src/test

Script	Description
audio_getter.js	Get audio stream from HTTP endpoint
client_stresser.js	Stress test the server by spamming request for audio data
ws_client.js	WebSocket client example

Stress Test Results 🧪

• Audio delay will change depending on client count
• Tested on Windows 10 Pro, Ryzen 5 3500X, NVIDIA GTX 1660 Super

Client side 💻

• Client Count is approximated

Client Count	Symptoms
0 - 1999	Not noticeable - None
1200	Small amount of short stutter
1500	Much longer and more frequent stutter
1800+	Very frequent stutter, audio alternation (speed up)

---

Server side 🔗

• Client Count is approximated
• CPU usage 18-20% (Not affected by client count)

Client Count	Memory Used (MB)
0 (START)	50
500	100
1000	165
1500	220
1750	310
2000	405
2200	425
2350	480
2500	520
0 (END)	110

License 📰

This project is using MIT license © Edqe14
Click here to read LICENSE.