ydls

ydls

灵活多功能的在线媒体处理平台

ydls项目提供了一个基于yt-dlp的HTTP服务,实现了多源媒体下载、流传输、转封装和转码功能。系统集成了丰富的格式和编解码器支持,可应用于媒体格式转换、播客制作和音频提取等领域。通过简洁而强大的API设计,开发者能够精确控制输出格式、编解码器选择和时间段截取,为各类媒体处理任务提供灵活解决方案。

yt-dlpHTTP服务媒体下载转码DockerGithub开源项目

youtube-dl HTTP service

HTTP service for yt-dlp that downloads media from requested URL and streams, transmux and transcode to requested format if needed.

It can be used to convert media, create podcasts and audio only versions of media from various site like youtube, vimeo etc.

Docker image uses ffmpeg from wader/static-ffmpeg which can handle most format and codecs. Default configuration can transcode to these format, container and codecs:

FormatContainerAudioVideoSubtitle
alacmp4alac
flacflacflac
gifgifgif
m4amp4aac
mp3mp3mp3
oggoggvorbis, opus
wavwavpcm_s16le
mkvmatroskaaac, alac, ac3, mp3, vorbis, opus, flach264, hevc, vp8, vp9, theora, av1subrip, ass
mp4mp4aac, alac, mp3, vorbis, flach264, vp9, av1, hevcmov_text
mxfmxfpcm_s16lempeg2video
tsmpegtsaac, mp3, ac3h264, hevc
webmwebmvorbis, opusvp8, av1, vp9webvtt
rssmp3mp3

The rss format transforms a playlist into a RSS audio podcast.

See ydls.json for more details.

Usage

Run with docker

Pull mwader/ydls or build image using the Dockerfile. Run a container and publish TCP port 8080 somehow.

docker run -p 8080:8080 mwader/ydls

Build and install yourself

Run go get github.com/wader/ydls/cmd/ydls to install ydls. Make sure you have ffmpeg, yt-dlp, rtmpdump and mplayer installed and in PATH.

Copy and edit ydls.json to match your ffmpeg builds supported formats and codecs.

Start with ydls -server -config /path/to/ydls.json and it default will listen on port 8080.

Endpoints

Download and make sure media is in specified format:
GET /<format>[+option+option...]/<URL-not-encoded>
GET /?format=<format>&url=<URL>[&codec=...&codec=...&retranscode=...]

Download in best format:
GET /<URL-not-encoded>
GET /?url=<URL-encoded>

Parameters

format - Format name. See table above and ydls.json
URL - Any URL that yt-dlp can handle
URL-not-encoded - Non-URL-encoded URL. The idea is to be able to simply prepend the download URL with the ydls URL by hand without doing any encoding (for example in the browser location bar)
codec - Codec to use instead of default for format (can be specified one or two times for audio and video codec)
retranscode - Retranscode even if input codec is same as output
time - Only download specificed time range. Ex: 30s, 20m30s, 1h20m30s will limit duration. 10s-30s will seek 10 seconds and stop at 30 seconds (20 second output duration)
items - If playlist only include this many items

option - Codec name, time range, retranscode or <N>items

Examples

Download and make sure media is in mp3 format:
http://ydls/mp3/https://www.youtube.com/watch?v=cF1zJYkBW4A

Download using query parameters and make sure media is in mp3 format:
http://ydls/?format=mp3&url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DcF1zJYkBW4A

Download and make sure media is in webm format:
http://ydls/webm/https://www.youtube.com/watch?v=cF1zJYkBW4A

Download and make sure media is in mkv format using mp3 and h264 codecs:
http://ydls/mkv+mp3+h264/https://www.youtube.com/watch?v=cF1zJYkBW4A

Download and retranscode to mp3 even if input is already mp3:
http://ydls/mp3+retranscode/https://www.youtube.com/watch?v=cF1zJYkBW4A

Download specified time range in mp3:
http://ydls/mp3+10s-30s/https://www.youtube.com/watch?v=cF1zJYkBW4A

Download in best format:
http://ydls/https://www.youtube.com/watch?v=cF1zJYkBW4A

Playlist as audio podcast with 3 latest items:
http://ydls/rss+3items/https://www.youtube.com/watch?list=PLtLJO5JKE5YCYgIdpJPxNzWxpMuUWgbVi

Tricks and known issues

For some formats the transcoded file might have zero length or duration as transcoding is done while streaming. This is usually not a problem for most players.

Download with curl and save to filename provided by response header:

curl -OJ http://ydls-host/mp3/https://www.youtube.com/watch?v=cF1zJYkBW4A

Docker image can download from command line. This will download in mp3 format to current directory:

docker run --rm -v "$PWD:$PWD" -w "$PWD" mwader/ydls https://www.youtube.com/watch?v=cF1zJYkBW4A mp3

yt-dlp URL can point to a plain media file.

If you run the service using some cloud services you might run into geo-restriction issues with some sites like youtube.

Development

When fiddling with ffmpeg and yt-dlp related code I usually do this:

docker build --target ydls-dev -t ydls-dev . && docker run --rm -ti -v "$PWD:/$PWD" -w "$PWD" ydls-dev

Then inside dev container:

# run cli version go run cmd/ydls/main.go -config ./ydls.json -debug # run all tests CONFIG="$PWD/ydls.json" TEST_EXTERNAL=1 go test -v -cover -race ./...

TODO

  • Optional stream for format? example mp4 at least video or audio?
  • Download playlist as zip archive.
  • JSON output?
  • Refactor messy downloadFormat function. Maybe a media interface with audio, video and subtitle implementations that select best format and return reader and mappings? should share a common format picker so formats can be shared and not re-downloaded.
  • Make download part into an interface to support more than yt-dlp.
  • Bitrate factor per codec when sorting formats (prefer aac over mp3 at same bitrate etc)
  • X-Remote IP header?

License

ydls is licensed under the MIT license. See LICENSE for the full license text.

编辑推荐精选

蛙蛙写作

蛙蛙写作

AI小说写作助手,一站式润色、改写、扩写

蛙蛙写作—国内先进的AI写作平台,涵盖小说、学术、社交媒体等多场景。提供续写、改写、润色等功能,助力创作者高效优化写作流程。界面简洁,功能全面,适合各类写作者提升内容品质和工作效率。

AI助手AI工具AI写作工具AI辅助写作蛙蛙写作学术助手办公助手营销助手
Trae

Trae

字节跳动发布的AI编程神器IDE

Trae是一种自适应的集成开发环境(IDE),通过自动化和多元协作改变开发流程。利用Trae,团队能够更快速、精确地编写和部署代码,从而提高编程效率和项目交付速度。Trae具备上下文感知和代码自动完成功能,是提升开发效率的理想工具。

热门AI工具生产力协作转型TraeAI IDE
问小白

问小白

全能AI智能助手,随时解答生活与工作的多样问题

问小白,由元石科技研发的AI智能助手,快速准确地解答各种生活和工作问题,包括但不限于搜索、规划和社交互动,帮助用户在日常生活中提高效率,轻松管理个人事务。

聊天机器人AI助手热门AI工具AI对话
Transly

Transly

实时语音翻译/同声传译工具

Transly是一个多场景的AI大语言模型驱动的同声传译、专业翻译助手,它拥有超精准的音频识别翻译能力,几乎零延迟的使用体验和支持多国语言可以让你带它走遍全球,无论你是留学生、商务人士、韩剧美剧爱好者,还是出国游玩、多国会议、跨国追星等等,都可以满足你所有需要同传的场景需求,线上线下通用,扫除语言障碍,让全世界的语言交流不再有国界。

讯飞智�文

讯飞智文

一键生成PPT和Word,让学习生活更轻松

讯飞智文是一个利用 AI 技术的项目,能够帮助用户生成 PPT 以及各类文档。无论是商业领域的市场分析报告、年度目标制定,还是学生群体的职业生涯规划、实习避坑指南,亦或是活动策划、旅游攻略等内容,它都能提供支持,帮助用户精准表达,轻松呈现各种信息。

热门AI工具AI办公办公工具讯飞智文AI在线生成PPTAI撰写助手多语种文档生成AI自动配图
讯飞星火

讯飞星火

深度推理能力全新升级,全面对标OpenAI o1

科大讯飞的星火大模型,支持语言理解、知识问答和文本创作等多功能,适用于多种文件和业务场景,提升办公和日常生活的效率。讯飞星火是一个提供丰富智能服务的平台,涵盖科技资讯、图像创作、写作辅助、编程解答、科研文献解读等功能,能为不同需求的用户提供便捷高效的帮助,助力用户轻松获取信息、解决问题,满足多样化使用场景。

模型训练热门AI工具内容创作智能问答AI开发讯飞星火大模型多语种支持智慧生活
Spark-TTS

Spark-TTS

一种基于大语言模型的高效单流解耦语音令牌文本到语音合成模型

Spark-TTS 是一个基于 PyTorch 的开源文本到语音合成项目,由多个知名机构联合参与。该项目提供了高效的 LLM(大语言模型)驱动的语音合成方案,支持语音克隆和语音创建功能,可通过命令行界面(CLI)和 Web UI 两种方式使用。用户可以根据需求调整语音的性别、音高、速度等参数,生成高质量的语音。该项目适用于多种场景,如有声读物制作、智能语音助手开发等。

咔片PPT

咔片PPT

AI助力,做PPT更简单!

咔片是一款轻量化在线演示设计工具,借助 AI 技术,实现从内容生成到智能设计的一站式 PPT 制作服务。支持多种文档格式导入生成 PPT,提供海量模板、智能美化、素材替换等功能,适用于销售、教师、学生等各类人群,能高效制作出高品质 PPT,满足不同场景演示需求。

讯飞绘文

讯飞绘文

选题、配图、成文,一站式创作,让内容运营更高效

讯飞绘文,一个AI集成平台,支持写作、选题、配图、排版和发布。高效生成适用于各类媒体的定制内容,加速品牌传播,提升内容营销效果。

AI助手热门AI工具AI创作AI辅助写作讯飞绘文内容运营个性化文章多平台分发
材料星

材料星

专业的AI公文写作平台,公文写作神器

AI 材料星,专业的 AI 公文写作辅助平台,为体制内工作人员提供高效的公文写作解决方案。拥有海量公文文库、9 大核心 AI 功能,支持 30 + 文稿类型生成,助力快速完成领导讲话、工作总结、述职报告等材料,提升办公效率,是体制打工人的得力写作神器。

下拉加载更多