
yt-dlp Video Link Extractor - Any URL to Links, 1000+ Sites
io.github.AnshumanAtrey由 GitHub 帳號 AnshumanAtrey 發布v0.1.6更新於 Oct 29, 2026
yt-dlp Link Extractor - Any Video URL to Direct Links
概覽
託管於 Apify 的 MCP 伺服器,可將 1000 多個 yt-dlp 網站的影片網址解析成直連連結、字幕與中繼資料。
- 可以做什麼
- 將一個或多個影片、播放清單或頻道網址解析成直連媒體連結
- 回傳最佳合併連結,以及分開的最佳純視訊與純音訊串流網址
- 列出所有可用格式及其各自的網址、解析度與編碼
- 回傳標題、上傳者、長度、觀看次數、上傳日期等中繼資料
- 回傳縮圖與依語言分類的字幕/自動字幕網址
- 適用情境
- 當助理需要取得 YouTube、Vimeo、SoundCloud、Twitch 等網站影片的可播放或可下載直連連結,或需要批次取得影片中繼資料,或想把串流網址交給自己的下載器、轉檔器或播放器時適用。它只回傳連結,並非檔案下載工具。
- 執行需求
- 遠端 MCP 端點 mcp.apify.com,使用 OAuth 驗證;需要 Apify API 權杖,以 Authorization 標頭中的 Bearer 形式送出。執行費用計入該 Apify 帳號。選用輸入包括代理設定,以及用於年齡限制或需登入影片的 Netscape 格式 cookies。
安裝
在 SourceWeft 中
- 開啟 儀表板中的 yt-dlp Video Link Extractor - Any URL to Links, 1000+ Sites,將其新增到工作區。
- 連接 所需的 憑證。它們會被私密儲存,絕不會在公開頁面上展示。
- 為需要使用其工具的對話啟用該服務。
Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。
交給代理程式安裝
把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式,或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南,先向你展示發布者、端點和登入方式,經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。
閱讀 https://sourceweft.com/mcp/io.github.AnshumanAtrey%2Fyt-dlp-video-link-extractor/install.md,並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前,先告訴我它的發布者、端點和登入方式,等我確認。不要把 API 金鑰寫進命令或我會提交的檔案;問我金鑰應從哪裡取得。其他 MCP 客戶端
把它新增到你客戶端的 mcpServers 設定中。 客戶端會在首次連接時完成 OAuth 登入。
{
"mcpServers": {
"yt-dlp-video-link-extractor-any-url-to-l": {
"type": "http",
"url": "https://mcp.apify.com/?tools=anshumanatrey/yt-dlp-video-link-extractor"
}
}
}README
yt-dlp Video Link Extractor - Any URL to Links, 1000+ Sites
Paste any video URL and get back the direct stream and download links plus full metadata, resolved by yt-dlp. No files are stored, no RAM is burned - it resolves the links, it does not host the bytes.
Available as an Apify Actor. $0.002 per run + $0.003 per URL processed + $0.04 per video resolved.
What does it do?
Give it one or more video URLs. For each one it returns the direct media links (a single combined audio+video file when one exists, plus the best video-only and audio-only streams), every available format with its own URL, thumbnails, subtitles, and the metadata: title, uploader, duration, view count, upload date.
It is powered by yt-dlp, so it works on 1,872 sites (verified at build time) - YouTube, Vimeo, SoundCloud, Twitch, Dailymotion, Reddit, Rumble, Bilibili and far more - not just YouTube. You paste a URL, you get the links. That is the whole tool.
It never downloads the actual file. That is deliberate: downloading large videos into cloud storage is slow, expensive, and memory-heavy. Resolving the link is instant and lets you fetch, stream, or transcode the media yourself.
How is it different from a YouTube downloader?
Most actors in this space stop at YouTube. This one hands you the raw links for almost any video site, so you can build your own download, archive, or analysis pipeline on top.
When should I use it?
- Get a direct, playable/downloadable link for a video on almost any site
- Pull metadata (title, duration, views, upload date, uploader) in bulk
- Feed stream URLs into your own downloader, transcoder, or player
- Grab subtitle and caption track URLs across languages
- Resolve every video in a playlist or channel to its links in one run
How much does it cost to resolve a video?
Pay-per-event:
- $0.002 per run
- $0.003 per URL processed (covers the resolve attempt for any URL)
- $0.04 per video successfully resolved
A single video costs about $0.045. A failed URL costs only $0.005 (run + URL fee, never the $0.04 resolution fee). Resolving a 50-video playlist costs about $2.00. On paid Apify plans the URL and video fees are 10% less on Scale ($0.036 a video) and 20% less on Business and above ($0.032).
If you set a maximum cost per run, the actor stops when it is reached, never starts a URL it could not deliver, and says so in the run's status message.
Which inputs does it take?
What does the output look like?
One record per resolved video. The link-first fields:
best_urlis the single link most people want: a combined audio+video file when the site offers one, otherwise the best audio (for music/podcasts) or best video.best_video_url+best_audio_urlmatter for high-resolution sources. YouTube and Vimeo serve 1080p/4K video and audio as separate streams, sobest_url(the combined file) tops out lower - use the two separate links for maximum quality.- A failed URL returns
{ "input_url": ..., "ok": false, "error": "..." }.
How reliable is it?
yt-dlp is the most actively maintained extractor project in the world, and this actor follows its releases. A daily job builds each new yt-dlp release, runs it next to the current build on the same test videos, and ships it only if it resolves them at least as well, so extractor fixes usually reach your runs within a day and a release that does worse is held back. YouTube results include the combined audio+video link, because the actor ships the Deno JavaScript runtime that yt-dlp needs to solve YouTube's player challenges. When YouTube refuses a video to the cloud IP with a bot check, the actor retries that video through a residential proxy on its own, and vimeo.com links go through Vimeo's public player page, so neither needs your account.
Common questions
Q: Does it download the video file? No. It returns the direct links. Fetch, stream, or transcode the media yourself. This keeps it fast and cheap.
Q: Do the links expire? Often yes. Many sites (YouTube especially) issue time-limited URLs that expire in a few hours. Use them promptly.
Q: Which sites are supported? Anything yt-dlp supports - 1,872 extractors at last build, including YouTube, TikTok, Vimeo, Twitch, SoundCloud, Reddit, Dailymotion, Rumble, Bilibili, Facebook and more.
Q: A TikTok / Instagram / X link failed. Why? Those platforms aggressively block datacenter IPs. Supply a residential proxy in proxyConfiguration and/or cookies and retry. Most other sites work with no proxy.
Q: Does YouTube work? Yes: single videos, lists and playlists. YouTube answers some videos from cloud (datacenter) IPs with a "confirm you're not a bot" check. The actor then retries that video through a residential proxy automatically, at no extra charge, and the run's status message counts the retries. A retried video takes longer, about 20 to 40 seconds. Age-restricted and members-only videos still need cookies.
Q: Does Vimeo work? Yes. vimeo.com links, unlisted ones included, are resolved through Vimeo's public player page, which needs no account. View counts and upload dates come only with a Vimeo login: pass cookies for those. A video whose owner blocks embedding cannot be resolved without an account.
Q: Can I pass a whole playlist or channel? Yes. It expands to the videos inside, capped by playlistLimit.
Q: Can I get just the audio? Yes - best_audio_url is the best audio-only stream. For music sites it is also the best_url.
Q: Am I charged for failed URLs? Only the small per-URL processing fee ($0.003), which covers the resolve attempt. You are never charged the $0.04 resolution fee unless a video is actually resolved.
Limitations
- Returns links, not downloaded files (by design).
- Direct links are time-limited by the source platform - use them quickly.
- Datacenter-blocked sites (TikTok, Instagram, X) need a residential proxy and/or cookies.
- Live streams resolve to HLS/DASH manifest URLs, not a single finished file.
Ethical use
Download and use only content you have the right to access. Respect each platform's terms of service and copyright. This tool resolves publicly reachable media links; it does not bypass paywalls or DRM.
Other actors by the same maintainer
Part of a portfolio of cloud tools built and maintained by Anshuman Atrey (@AnshumanAtrey). Custom features shipped within 24-48h for legitimate use cases via atrey.dev.
License
MIT (see LICENSE). Built on top of yt-dlp (Unlicense / public domain).
來源:README.md,提交 07854f6
工具
0版本歷史
2- v0.1.6最新Oct 29, 2026
- v0.1.1Sep 24, 2026
評分
已滿足適用檢查項的 75%
該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。
- 近 30 天內可連通必需25 分
- 已宣告認證方式必需10 分
- 已列出工具15 分
請讓伺服器在無憑證時回應 tools/list,或發布可在沙箱中無憑證啟動的軟體包。
- 描述必需15 分
- 圖示10 分
- 文件15 分
- 授權條款10 分
請在儲存庫或註冊中心條目中宣告授權條款。