msojocs/bilibili-linux

[!NOTE] This file is generated by AI.

Simplified Chinese | Traditional Chinese | English

Introduction

This project involves:

1. Deobfuscation
2. Debugging
3. Brains

This project is completely open source, with no code encryption. If you have any doubts, please review the code yourself or stop using the related files.

Features

1. No annoying title bar
2. Full-screen support for low-resolution screens
3. Update check support
4. A menu to close the application
5. Roaming support
6. Search in other regions
7. Danmaku Sharing (Content in Chinese)

How to Use

Automated Builds

Please download from the Release page:

https://github.com/msojocs/bilibili-linux/releases

For continuous builds:

https://github.com/msojocs/bilibili-linux/releases/tag/continuous

Manual Build

1. Clone the code

   git clone https://github.com/msojocs/bilibili-linux.git
   cd bilibili-linux
   

2. Install

   tools/setup-bilibili.sh
   

3. Launch

   bin/bilibili
   

AppImage Installation

Please use AppImageLauncher

Gentoo Linux

Emerge from gentoo-zh overlay.

Additional Notes

If you don't want to install another Electron on your computer, you can extract app.asar from the release version and run it with your already installed Electron. The recommended Electron version is: 43.1.1 (VaapiOnNvidiaGPUs requires Electron 43+ / Chromium 130+).

GPU Acceleration Configuration

Electron 43 (Chromium 150) supports VA-API hardware video decoding on Linux. Below are the recommended configurations:

NVIDIA GPU

1. Install the VA-API → NVDEC backend:

   # Arch Linux
   sudo pacman -S libva-nvidia-driver
   # Debian/Ubuntu
   sudo apt install libva-nvidia-driver
   

2. Set the environment variable (libva needs this to find the NVIDIA backend):

   export LIBVA_DRIVER_NAME=nvidia
   

3. Configure in ~/.config/bilibili/bilibili-flags.conf:

   --ignore-gpu-blocklist
   --enable-features=AcceleratedVideoDecodeLinuxGL,AcceleratedVideoDecodeLinuxZeroCopyGL,VaapiOnNvidiaGPUs,VaapiIgnoreDriverChecks,PlatformHEVCDecoderSupport
   --enable-gpu-rasterization
   --enable-zero-copy
   --disable-gpu-sandbox
   

   Notes:

   • VaapiOnNvidiaGPUs (Chromium 130+) bypasses vaapi_wrapper's NVIDIA device skip check
   • AcceleratedVideoDecodeLinuxGL / AcceleratedVideoDecodeLinuxZeroCopyGL are the recommended GL zero-copy decode paths for Chromium 150
   • Electron 28 lacks these features — NVIDIA hardware video decode requires Electron 43+

Intel / AMD GPU

--ignore-gpu-blocklist
--enable-features=AcceleratedVideoDecodeLinuxGL,AcceleratedVideoDecodeLinuxZeroCopyGL,VaapiVideoDecoder,VaapiIgnoreDriverChecks,PlatformHEVCDecoderSupport
--enable-gpu-rasterization
--enable-zero-copy
--disable-gpu-sandbox

Intel iGPUs and AMD GPUs use the Mesa VA-API backend — no additional driver or environment variables needed.

Verifying Hardware Decode

After playing a video, check if the GPU decode engine is active:

cat /proc/$(pgrep -f "type=gpu-process.*bilibili")/fdinfo/* | grep drm-engine-video

A non-zero value indicates the GPU decode engine is working. You can also visit chrome://gpu to check the "Video Decode" status.

Note

• --use-gl=angle --use-angle=gl may break VA-API decoding on some Linux desktop environments (especially Wayland). The default EGL/GLX backend is recommended.
• If using a system proxy (e.g. Clash), bilipc.bilibili.com may fail to resolve through the proxy, causing a blank white screen. Electron 43+ includes built-in proxy bypass handling.

Developer Tools

F12 Developer Panel

You can now open the developer tools yourself:

Try pressing F12. If there's no response, it's not enabled.

The login interface has two layers: right-click to open the inner layer, and press F12 for the outer layer.

About Loongson

https://areweloongyet.com/docs/loong-or-loongarch

Distribution	Architecture Identifier
AOSC OS	loongarch64
Debian	Old World: loongarch64 New World: loong64
Gentoo	loong
Loong Arch Linux	loong64
RPM-based	loongarch64
Slackware	loong64

SponsorBlock Assistant

Automatically helps you jump to key moments.

Option	Description
Whisper Proxy	A proxy is needed for model downloads, e.g., http://127.0.0.1:1080
LD_LIBRARY_PATH	You might encounter a cudnn library not found issue, refer here
AI Recognition TOKEN	Token from https://www.bigmodel.cn/, using the free model glm-4.5-flash

Switch Language

1. Click the settings button in the bottom right corner of the homepage.
2. Go to "Other Settings".
3. Select your desired language in the "Language Settings" area.

Changelog

CHANGELOG.md

Preview

Thanks to the following projects

1. BiliRoaming
2. bilibili-helper
3. BilibiliSponsorBlock

Disclaimer

The license applies only to this project, not its products.

The copyright of the Bilibili client belongs to Shanghai Kuanyu Digital Technology Co., Ltd.

The maintainer is not responsible for any additional issues arising from the use of this project, such as account bans or theft. Please use with caution.

If there are any improprieties, please contact me at: [email protected]