hua0512 / stream-rec

Automatic streaming record tool using Kotlin coroutines, flow, serialization, sqldelight, and ktor. 虎牙/抖音//斗鱼/Twitch直播，弹幕自动录制

English | 简体中文

Stream-rec

Stream-rec is an automatic stream recording tool for various streaming services.

It's powered by Kotlin, Ktor, and ffmpeg.

This project is the result of my personal need for a tool that can automatically record live streams and upload them to cloud storage.

Note

This project is the result of my personal learning of Kotlin Coroutines, flow, Ktor, dao, repository pattern and other technologies.

Features

• Automatic stream recording, with configurable quality and format.
• Automatic file naming based on the stream title and start time.
• Automatic Danmu(Bullet comments) recording
• Persistent storage of stream and upload information (using SQLite)
• Integration with Rclone for uploading to cloud storage
• Configurable via web interface.
• Docker support

Supported streaming services

Service	Recording	Danmu
Huya	✅	✅
Douyin	✅	✅
Douyu	✅	✅
Bilibili	❌	❌
Twitch	❌	❌
Youtube	❌	❌
Niconico	❌	❌
AfreecaTv	❌	❌

• As for now, only Huya, Douyin, Douyu are supported. (Because I only use these three services 😄).
• More services will be supported in the future (if I have time, PRs are welcomed).

Screenshots

Installation

1. Docker Compose (Recommended)

1.1 Configuring the docker-compose.yml file

Create a docker-compose.yml, take a look at the example configuration file.

Please read the comments in the configuration file for more information and make sure to modify crucial parts like passwords, paths, etc.

1.2 Running the Docker compose

Make sure you are in the same directory as the docker-compose.yml file, then run the following command:

docker compose up -d && docker compose logs -f

Now, you are all set! You can access the web interface at http://localhost:15275 and start configuring the tool ( see Configuration).

Note

You can detach from the logs by pressing Ctrl + C. And you can reattach to the logs by running docker compose logs -f. To stop the containers, run docker compose down.

2. Building from source

2.1 Prerequisites

• Internet access, obviously 😂
• Git (optional, for cloning the repository)
• A java development kit (JDK) (version 17 or later), Amazon Corretto 17 is recommended.
• FFmpeg (Make sure it's in your PATH).
• Streamlink (optional, for recording streams, make sure it's in your PATH)
• Rclone (optional, for uploading to cloud storage, make sure it's in your PATH)
• Sqlite3 (for storing stream, upload information, make sure it's in your PATH)

2.2 Building the backend

To build the project, first clone the repository and navigate to the root directory of the project.

git clone https://github.com/hua0512/stream-rec.git
cd stream-rec

Then, build the project using the following command:

./gradlew stream-rec:build -x test

The built fat jar file stream-rec.jar will be located in the stream-rec/build/libs directory.

2.2.1 Running the jar file

To run the jar file, use the following command:

java -jar stream-rec/build/libs/stream-rec.jar

Several environment variables can be set to configure the tool:

• DB_PATH: Path to the SQLite database folder. (default: ./db)
• JWT_SECRET: Secret key for JWT token generation.
• LOG_LEVEL: Log level (default: info).
• LOGIN_SECRET: Login password for the web interface (default: stream-rec).

For example:

java -DDB_PATH=/path/to/your/db -DLOG_LEVEL=DEBUG -DJWT_SECRET=SECRET -DLOGIN_SECRET=123 -jar stream-rec/build/libs/stream-rec.jar

[!IMPORTANT]\

• Please set the LOGIN_SECRET environment variable to a secure password. This password is used to log in to the web interface.
• This password cannot be changed after the first run.

2.3 Building the frontend

Frontend is used to configure the tool, it's a simple web interface built with React.

Navigate to frontend repository and follow the build instructions.

After that, you can start configuring the tool by accessing the web interface at http://localhost:15275. Take a look at the Configuration page for more information.

Troubleshooting

Check logs under logs directory, by default, it will be created in the same directory as the DB_PATH.

• There´s a environment variable LOG_LEVEL that can be set to debug to enable debug logs.

Contributing

Contributions are welcome! If you have any ideas, suggestions, or bug reports, please feel free to open an issue or a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for details.

stream-rec's Issues

danmu end of line (</root>) is not properly appended sometimes

Docker support and fat jar

Finish on parted/streaming ended callbacks

doc: update docs

Disable protobuf plugin and use generated code

Protobuf gradle plugin takes too much time. Find a way to turn it off, kt serialization for protobuf is a good idea

Support other upload platforms

build error cause missing streamlink package in alpine linux

Wait for upstream fix

refactor: make live updates socket to send list of updates and add heart beat mechanism

closed by #36

docker 版的一些问题

• 时区不是**时区
• docker-compose.yml 文件没有映射 rclone 的配置目录，每次使用docker-compose up -d启动时 rclone 配置都会丢失

更新后处理模版会清除录播文件

会清除正在录播的文件

Custom content serialization to avoid class descriminators

Frontend web

Douyin danmu

Finish rclone upload

docker: configure actions to publish docker images

refactor: leverage segmentation to ffmpeg's segmenter

Notes:

• Using ffmpeg's segmenter should/will reduce platform's api calls.
• Only time segmentation is supported doc.
• File size and bitrate is not updated in real time

feat: streamlink support

Persistence storage