Downly Chrome Extension - Download Helper
除第 3 节所述的匿名使用统计外, 本扩展不向任何远端服务器发送数据. 在用户主动操作时, 扩展会读取以下数据, 仅用于在本机内传递给配套桌面客户端:
以下数据从不收集:
本扩展的单一目的是: 在支持的视频网站上识别可下载的视频, 帮助用户保存到本机. 下载有两种方式:
扩展在用户浏览器内进行的处理仅限以下范围, 全部在本地完成:
| 数据 | 用途 | 范围 |
|---|---|---|
| 视频流元数据 | 识别白名单站点正在播放的视频, 在扩展弹窗展示给用户选 | 仅当前 tab, chrome.storage.session 浏览器关闭自动清空 |
| 视频数据 (浏览器内下载) | 记录页面播放的音视频数据, 在本机合并成 MP4 后保存到用户选择的位置 | 合并期间暂存在浏览器的扩展私有存储中, 保存后删除; 中断遗留的暂存文件 24 小时后自动清理 |
| 右键链接 URL | 用户右键 "用 Downly 下载" 时拿到链接 + 当前页 URL, 转交本机桌面客户端 | 仅用户主动点击时, 不持久化 |
| 登录 cookie | 仅在用户点击下载时读取, 用于让站点服务器返回登录态内容 (高清画质 / 反 LOGIN_REQUIRED) | 白名单站点 (youtube.com / google.com), 严格按 cookie 名白名单; 下载链接 / 接管下载时为该链接本身的 cookie. 扩展不持久化 |
为了解有多少人在用、下载为何失败并据此改进, 扩展默认发送匿名使用统计, 可随时关闭. 除此之外不与任何第三方共享数据, 不集成广告服务.
posthog.com). 服务端根据请求 IP 推导国家/地区, 不保存原始 IP, 也不建立用户画像.本扩展请求 cookies 权限是为了在用户主动点击下载时, 读取白名单站点的登录 cookie, 一次性传给本机桌面客户端, 让站点服务器返回用户已登录状态下能访问的内容. 具体范围:
youtube.com / google.com: 读 yt-dlp 抓 YouTube 实际需要的关键 cookie (SAPISID / SID 等), 反 LOGIN_REQUIRED. 严格按 cookie 名白名单读取 (例如 youtube/google 域大约 100 个 cookie, 仅取 yt-dlp 抓 youtube 需要的约 15 个), 不读取分析 / 广告 / 其他服务的 cookie.chrome.cookies.getAll({ url })), 连同浏览器 User-Agent 一起交给本机桌面客户端, 用于下载需要登录才能访问的文件. 桌面客户端只把这些 cookie 发给与之匹配的网站 (按域名 / 路径匹配, 跳转到其他网站时不携带), 保存在本机任务记录中直到下载完成或任务被删除, 随后清除.| 权限 | 用途 |
|---|---|
webRequest | 在白名单站点上只读监听网络请求, 识别可下载的媒体流 URL (HLS / DASH / MP4 等). 不修改, 不拦截, 不重定向 |
downloads | 覆盖 blob URL 的默认 UUID 文件名为可读标题, 可选展示下载进度 |
storage | 用 chrome.storage.session 临时缓存当前 tab 嗅探结果, 浏览器关闭自动清空; 用 chrome.storage.local 保存设置 (如悬浮按钮位置、统计开关) 与尚未发送的匿名统计 |
tabs | 读取当前 tab URL 关联视频; 监听 tab 关闭事件清缓存 |
scripting | 仅在用户操作扩展弹窗时, 在当前 tab 注入小脚本读取 og:image / video.poster 用作下载 UI 封面 |
cookies | 见上述 Cookie 用途说明 |
declarativeNetRequestWithHostAccess | 为白名单站点的媒体 CDN 请求重写 Referer 头, 用于绕过防盗链 (否则下载返 HTTP 403) |
nativeMessaging | 跟用户本机的 Downly 桌面客户端通信. 通信仅限本机, 不走互联网 |
contextMenus | 在链接的右键菜单加一项 "用 Downly 下载", 仅在 link 上下文显示, 不在页面 / 选区 / 图片等其他上下文显示 |
host_permissions <all_urls> | 白名单站点的视频 CDN 域名按会话动态调度 (跟页面域名不同), webRequest 监听需观察跨域请求才能捕获实际媒体 URL; 非白名单页面发起的请求被静默丢弃, 不入库不上传 |
Downly 桌面客户端是配套的本机程序, 用户需单独下载安装. 它接收扩展转交的下载任务, 跟视频源服务器直接通信完成下载. 桌面客户端不向 Downly 项目方上传任何用户数据.
chrome://extensions/ 移除本扩展, 所有本地缓存立即失效本政策若有更新, 将通过版本说明公告. 用户继续使用即视为同意更新.
Apart from the anonymous usage statistics described in Section 3, this extension does not send any data to remote servers. When the user actively interacts with the extension, the following data is read solely for forwarding to the locally-installed companion desktop client:
The following data is never collected:
The single purpose of this extension is: detect downloadable videos on supported video sites and help the user save them to their device. There are two ways to download:
The extension's processing is limited to the following, all performed locally inside the browser:
| Data | Purpose | Scope |
|---|---|---|
| Video stream metadata | Identify videos playing on whitelisted sites; display in the popup for user selection | Only the active tab; chrome.storage.session is automatically cleared on browser close |
| Video data (in-browser download) | Record the audio and video data the page plays, merge it into an MP4 locally, and save it to the location the user chooses | Kept in the extension's private browser storage only while merging and deleted after saving; leftovers from interrupted downloads are cleaned up after 24 hours |
| Right-click link URL | When user picks "Download with Downly" on a link, capture the link URL + current page URL, forward to local companion | Only when user actively clicks; not persisted |
| Login cookies | Read only when user clicks download, used so the source server returns content the user is entitled to in their logged-in state (HD streams / bypass LOGIN_REQUIRED) | Allowlisted sites (youtube.com / google.com) with a strict cookie-name allowlist; for download links / takeover, the cookies of that link itself. Not persisted by the extension |
To understand how many people use the extension and why downloads fail, the extension sends anonymous usage statistics by default; you can turn this off at any time. Beyond that, no data is shared with any third party, and no advertising service is integrated.
posthog.com). The server derives a country/region from the request IP and does not store the IP address; no user profiles are created.This extension requests the cookies permission so that, when the user actively clicks download, it can read login cookies from a small allowlist of source sites and forward them once to the local companion desktop client, so that the source server returns content the user is entitled to in their logged-in state. Specific scope:
youtube.com / google.com: reads only the cookies yt-dlp actually requires to fetch YouTube content (SAPISID / SID etc.), bypassing LOGIN_REQUIRED. Cookies are read using a strict per-domain name allowlist (the youtube/google domains have ~100 cookies in total but we only read ~15 needed by yt-dlp). Analytics / advertising / other-service cookies are never read.chrome.cookies.getAll({ url })) and passes them, together with the browser User-Agent, to the local desktop client so files that require login can be downloaded. The desktop client sends these cookies only to the matching site (matched by domain and path, never carried over to a different site after a redirect), keeps them in its local task record until the download finishes or the task is deleted, and then clears them.| Permission | Use |
|---|---|
webRequest | Read-only monitoring of network requests on whitelisted sites to detect downloadable media stream URLs (HLS / DASH / MP4 etc.). Does not modify, block, or redirect. |
downloads | Override blob URL default UUID filenames with human-readable titles; optionally surface download progress. |
storage | Cache per-tab sniffed media metadata in chrome.storage.session, automatically cleared on browser close; keep settings (e.g. floating button position, the statistics switch) and not-yet-sent anonymous statistics in chrome.storage.local. |
tabs | Read the active tab URL to associate detected media with the correct page; listen to tab close events to clean up cache. |
scripting | On user popup interaction only, inject a tiny snippet into the active tab to read og:image / video.poster as a download UI thumbnail. Not used for general DOM injection. |
cookies | See "Cookie Permission" above. |
declarativeNetRequestWithHostAccess | Rewrite Referer header for media CDN requests on whitelisted sites, required to bypass anti-hotlinking that would otherwise return HTTP 403. |
nativeMessaging | Communicate with the user's locally-installed Downly desktop client. Local communication only, never via the internet. |
contextMenus | Add a single "Download with Downly" item to the link right-click menu. Shown only on link contexts, never on pages / selections / images / other contexts. |
host_permissions <all_urls> | Allowlisted sites stream media through CDN domains chosen dynamically per session (separate from the page hostname); webRequest needs cross-origin visibility to capture the actual media URL. Requests originating from non-allowlisted pages are silently dropped — never logged or uploaded. |
The Downly Desktop Client is a companion native application that the user installs separately. It receives download tasks from the extension and communicates directly with video source servers to perform the download. The desktop client does not upload any user data to the Downly project.
chrome://extensions/; local cache is destroyed immediately.Updates to this policy will be announced via release notes. Continued use indicates acceptance.