PyMKUI

PyMKUI is a modern web management interface for ZLMediaKit. Built on deep integration through the Python plugin mechanism, it provides an intuitive, polished way to manage your streaming media server.

⚠️ The project is under rapid development and the database schema changes frequently. After every code update, delete the old database first:

rm data/pymkui.db

The database is rebuilt automatically on the next start.

---

Features

• 🎬 Stream management — view, play, and stop streams; capture snapshots
• 📡 Pull-stream proxy — multiple backup sources, on-demand/immediate modes, automatic failover, persistent recovery, online editing
• 🔄 Transcode presets — save multiple protocol-conversion parameter sets and load them with one click when pulling a stream; supports loading the server defaults
• 👥 Viewer list — see online viewers and connection details for each stream in real time
• 📊 Server monitoring — real-time charts for CPU, memory, disk, and network
• ⚙️ Service configuration — read and write ZLMediaKit configuration items online
• 🌐 Browser publishing — WHIP-based live publishing straight from the browser
• 🔗 Network connections — view and manage all current TCP/UDP sessions
• 📁 Recording management — recordings organized by stream and by date, with search, in-browser playback and download, and automatic policy-based cleanup of expired files
• 🧩 Plugin system — built-in extensible Python event hooks; implement custom business logic such as publish authentication, recording callbacks, and stream on/offline notifications without modifying the core code
• 🩺 Video quality probe — real-time sampling and analysis of published/pulled streams, reporting multi-dimensional quality metrics (bitrate, frame rate, GOP, audio/video interleaving, and more) as line and scatter charts to quickly pinpoint anomalies such as stuttering and artifacts

---

Quick start (one-command Docker deployment)

Already integrated into the ZLMediaKit Docker image — no manual configuration required:

docker run -id \
  -p 1935:1935 \
  -p 80:80 \
  -p 443:443 \
  -p 554:554 \
  -p 10000:10000 \
  -p 10000:10000/udp \
  -p 8000:8000/udp \
  -p 9000:9000/udp \
  zlmediakit/zlmediakit:master_py

Once it starts, open http://<server-IP>/ in your browser and log in with your api.secret key.

Ports

Port	Protocol	Purpose
80	TCP	HTTP (frontend + API)
443	TCP	HTTPS / WSS
1935	TCP	RTMP
554	TCP	RTSP
10000	TCP/UDP	RTP
8000	UDP	WebRTC
9000	UDP	SRT

---

Manual deployment (from source)

For cases where you compile ZLMediaKit yourself and want to use PyMKUI.

Project structure

pymkui/
├─ frontend/          # Static frontend pages
├─ backend/           # Python plugin and FastAPI interface
│  ├─ mk_plugin.py    # ZLMediaKit Python plugin entry point
│  ├─ py_http_api.py  # FastAPI HTTP API
│  ├─ database.py     # SQLite database
│  ├─ config.py       # Path configuration
│  ├─ mk_logger.py    # Logging wrapper
│  └─ shared_loop.py  # Shared asyncio event loop
├─ data/              # Generated automatically at runtime
│  └─ pymkui.db       # ⚠️ Delete when upgrading
└─ README.md

Step 1: Install the Python dependencies

Requires Python 3.10+

cd pymkui/backend
pip install -r requirements.txt

conda / venv

# conda
conda create -n pymkui python=3.12 && conda activate pymkui
pip install -r requirements.txt

# venv (Linux/Mac)
python -m venv venv && source venv/bin/activate && pip install -r requirements.txt

# venv (Windows)
python -m venv venv && venv\Scripts\activate && pip install -r requirements.txt

Step 2: Enable Python support when building ZLMediaKit

# Install the Python-related dependencies
apt-get update && apt-get install -y python3 python3-dev python3-pip
# Enable the Python features during cmake
cmake .. -DENABLE_PYTHON=ON

Step 3: Edit config.ini

[python]
plugin=mk_plugin

[http]
rootPath=/path/to/pymkui/frontend

Step 4: Let ZLMediaKit find the backend module

# Linux/Mac
export PYTHONPATH=/path/to/pymkui/backend:$PYTHONPATH

# Windows
set PYTHONPATH=C:\pymkui\backend;%PYTHONPATH%

Step 5: Start and verify

# Start ZLMediaKit
./MediaServer

# Verify that the frontend is reachable
curl -sv http://localhost:80/login.html
# Should return HTTP/1.1 200 OK

---

FAQ

Issue	Solution
ZLM prints red Python error logs	Update both ZLM and PyMKUI to the latest versions
/index/pyapi/* returns "not logged in"	Expected — the API is reachable; just complete the login
Still getting 404 after setting rootPath	Make sure the path points to the frontend/ directory and that login.html exists
Database error on startup	Delete data/pymkui.db and restart

---

Tech stack

Layer	Technology
Frontend	HTML5 + Tailwind CSS + Font Awesome
Player	Jessibuca (FLV), native HLS, WHEP (WebRTC)
Backend	Python + FastAPI (embedded in the ZLMediaKit Python plugin)
Database	SQLite
Server	ZLMediaKit (C++)

---

Roadmap

• 🤖 AI inference integration — integrate ONNX / TensorRT / OpenCV DNN for real-time video analysis (object detection, face recognition, and so on)

Contributing

Issues and pull requests are welcome.

License

MIT License