HeapLeach

Paste a heap of links. Download them together.

A bulk download manager for 180+ sites, with a web UI and a command-line mode. Expand albums, folders and playlists into a queue, download files in parallel, and resume interrupted transfers. Runs on Linux, macOS, Windows and Docker; the web UI is built into the executable.

Download · Quick start · Documentation · Supported sites · Contribute · Get help

Why HeapLeach?

• Batch downloads: paste one link per line; supported albums, folders and playlists expand automatically.
• Resumable transfers: keep partial files, retry failures, and continue after an interruption.
• Control your bandwidth: adjust parallel downloads, connections per file and the native download speed limit while the queue runs.
• A queue you can navigate: search, filter, pause, retry individual files, and inspect logs from the browser. Light, dark and mobile layouts included.
• Desktop or automation: open the local web UI, pass URLs to the CLI, or run the Docker image on your server.
• One application: Go backend and embedded React UI. Releases include local CAPTCHA model data; Docker also includes the media helpers.

Quick start

Download and run

1. Open the latest release and choose the archive for your system:

   System	Archive ending
   Linux, Intel / AMD 64-bit	linux_amd64.tar.gz
   Linux, ARM64	linux_arm64.tar.gz
   macOS, Apple silicon	darwin_arm64.tar.gz
   macOS, Intel	darwin_amd64.tar.gz
   Windows, Intel / AMD 64-bit	windows_amd64.zip

2. Extract the whole archive into a folder. Keep ocr/ beside the executable. No Go or Node.js installation is needed.

3. Run heapleach (heapleach.exe on Windows). It opens your browser and saves to your Downloads folder.

4. Paste your links, one per line, and click Add to queue.

Terminal commands for Linux, macOS and Windows

Run these from a folder containing only the archive version you want to use.

# Linux x86-64; use linux_arm64 for an ARM64 machine.
tar -xzf heapleach_*_linux_amd64.tar.gz
./heapleach

# macOS Apple silicon; use darwin_amd64 for an Intel Mac.
tar -xzf heapleach_*_darwin_arm64.tar.gz
./heapleach

# Windows PowerShell
Expand-Archive -Path .\heapleach_*_windows_amd64.zip -DestinationPath .\heapleach
.\heapleach\heapleach.exe

An argument-free launch uses a free local port. After you close the browser, it exits once downloads are idle and no tab has connected for a minute. Unfinished jobs return held after a restart; click Resume to continue.

See installation and updates for checksums, Docker persistence, source builds and platform notes.

Docker

On Linux or macOS with Docker, run:

mkdir -p "$HOME/Downloads"
docker run --rm -it --name heapleach \
  -p 127.0.0.1:8080:8080 \
  --user "$(id -u):$(id -g)" \
  -e HEAPLEACH_STATE=/downloads/.heapleach-state/queue.json.zst \
  -v "$HOME/Downloads:/downloads" \
  ghcr.io/johanlindvall/heapleach:latest

Open localhost:8080. Downloads and queue state stay in your Downloads folder when the container stops. The image supports amd64 and arm64 and includes yt-dlp, ffmpeg, ffprobe, deno and local OCR. latest follows successful main builds; use a versioned image to pin a manual release.

The web UI has no login. The command above publishes it only on your computer; see network access for remote deployments.

Usage

# Open the UI on a free local port and choose the destination.
./heapleach -addr 127.0.0.1:0 -open ~/Downloads

# Download a list of URLs, then exit. Blank lines and # comments are ignored.
./heapleach - ~/Downloads < urls.txt

# Limit native download speed and parallel transfers.
./heapleach -addr 127.0.0.1:8080 -concurrency 4 -max-speed 10MB ~/Downloads

# Show every command-line option.
./heapleach -help

urls.txt contains your own URLs, one per line. On PowerShell, use Get-Content .\urls.txt | .\heapleach.exe - to feed a list to the CLI. To pass URLs directly, quote each one, especially links containing & or #.

The CLI exits 0 on success, 1 if any input or download fails, and 130 after Ctrl-C. Partial files remain available for resume. Native transfers honor the shared speed limit; external helpers use their own controls.

Full usage guide · Configuration · Troubleshooting

Supported links

File hosts, image galleries, video sites, broadcasters, image boards and compatible platforms are supported, including:

Content	Examples
Files and folders	gofile, bunkr, MEGA, pixeldrain, MediaFire, Dropbox
Video and playlists	YouTube through yt-dlp, Vimeo, PeerTube
Broadcasts	SVT Play and other supported broadcasters
Galleries and collections	Chevereto, image boards, Internet Archive
Direct and embedded media	File URLs, HLS playlists (a DASH .mpd is refused with the reason), open directories, supported links embedded in webpages

Some workflows need optional helpers. Support depends on what a host makes available; see accepted URL formats and limitations. Sites change, so fixture tests alone cannot guarantee current live support.

Every supported site

Expand the complete site inventory, generated from the extractor registry

98 extractors, each chosen by what a page serves rather than by its domain, so a site that moves, or an install nobody listed, is still recognised.

Extractor	Recognises
4chan	4chan threads (/<board>/thread/<number>) whose page renders the thread's own t<number> container on the board and names the site
abclisten	episodes and programmes (/listen/programs/…) of ABC listen, by the page props its __NEXT_DATA__ island carries
alohatube	Aloha Tube view pages (/ktm/view.cgi?gid=<id>) that frame another site's player (an off-site <iframe allowfullscreen>), resolved where the video lives
archive.org	items (/details/<identifier>, /download/<identifier>[/<file>] and the other item routes) whose page is the archive's item page or file listing, or whose metadata API answers for the identifier
ard	videos (/video/…/<id>) and programmes (/sendung/, /serie/, /film/) of the ARD Mediathek, whose page loads from its page gateway
arte	programmes and collections (/<lang>/videos/<id>/, /embeds/<lang>/<id>) of ARTE, whose page calls its player API (/api/player/v2/)
autoindex	generated directory listings at any path: a page titled as an index of its own directory (Index of /…, Directory listing for /…) that links at least one entry inside it
balbums	searches of a Bunkr album index (/?search=<query>) whose page carries the index's search form, and its front page by that form and its "Page n of m" count, which is refused
bandcamp	releases (/album/<slug>, /track/<slug>), a band's /music and its front page, where the page is built from Bandcamp's client bundle and carries data-tralbum release data or the music grid
bandzoogle	pages of Bandzoogle band sites, at any path, whose player lists tracks as data-zoogle-track anchors
bbcsounds	episodes, brands and series (/sounds/play/<pid>, /sounds/brand/<pid>, /sounds/series/<pid>) served by the BBC Sounds web app
bilibili	videos (/video/BV<id>) and bangumi episodes whose page carries the bilibili player state (window.__INITIAL_STATE__ with its play information), short links included, and user spaces, series, lists and audio pages carrying its state or seed scripts
bitchute	videos and channels (/video/<id>/, /embed/<id>/, /channel/<id>/) served by the BitChute web app, a video naming its embed player
booru	image-board tag searches and posts on Danbooru, e621, Moebooru, Gelbooru, Philomena, Twibooru and szurubooru boards (/posts?tags=, /posts/<id>, /post/show/<id>, /index.php?page=post, /images/<id>, /post/<id>) whose page carries the software's markup and whose API answers for the same post or listing; Gelbooru 0.1 boards by their listing markup alone
bunkr	file pages (/f/<slug>, /v/<slug>, /i/<slug>) carrying a numeric data-file-id, and albums (/a/<slug>) listing theItem cards that link to file pages, or named Bunkr and linking file pages
celeb	creator galleries and images (/creator/<name>/images[/<id>], /creator/<name>) whose page is the site's Inertia app carrying the gallery's Creator/Media props in data-page
chevereto	Chevereto items, albums and user listings: pages below the site root carrying the platform's CHV.obj script namespace, or its generator or credit beside the item viewer, the data-object listing metadata or the password gate
civitai	images, posts, models and user galleries (/images/<id>, /posts/<id>, /models/<id>, /user/<name>) whose site's public API answers for that item on the page's own origin
coomerfans	creator pages (/u/<service>/<id>/<name>) listing that creator's /p/ posts, and posts (/p/<post>/<id>/<service>) carrying the post-body the media is read from
cum	creators and posts (/creators/<service>/<id>[/post/<post>]) whose page advertises its media servers as EDGE_BASE_URLS in the app's configuration
cyberdrop	albums (/a/<id>) listing #file entries with a file-size, and file pages (/f/<slug>, /e/<slug>) whose script asks the /api/file/info/ metadata endpoint
dailymotion	videos, playlists and user pages (/video/<xid>, /playlist/<xid>, /<user>) rendered by the Dailymotion server-side app (its neon bundle), short links included, and its embeddable player (/player/<id>.html?video=<xid>, the player id optional)
direct	anything else: a page's supported links or embedded player, or the URL as a plain file
doodstream	players (/e/<code>) naming the DoodStream /pass_md5/ token endpoint, landing pages (/d/<code>) naming DoodStream in their title or site metadata and framing its player, or either form showing its player gate, exact bare address refusal or removed-video page
drive	file links (/file/d/<id>, /open?id=, /uc?id=) whose page is the Google Drive file viewer (its apps-fileview bundle), the download confirmation form or Drive's quota or access refusal; folder links (/drive/folders/<id>) named Google Drive, which are refused
dropbox	share links (/s/…, /sh/…, /scl/fi/…, /scl/fo/…) whose page is the Dropbox viewer, loading its /static/atlas/ bundle
drtuber	watch pages and players (/video/<id>/<slug>, /embed/<id>) whose player is configured from the DrTuber player_config_json endpoint
drtv	programmes, seasons and series (/drtv/…/<slug>_<id>) of DR TV, served by its app with the state it boots from; audio pages (/lyd/…, /radio/…) served by the broadcaster's own front end are handed to yt-dlp
emload	file shares (/v2/file/<id>/<name>) whose page carries the dl-file download panel, or its removed-file or password notice
eporner	videos (/video-<id>/<slug>/, /hd-porn/<id>/) whose page sets the EPORNER player's video id and hash, and profile video tabs rendered with its profile script
erome	albums (/a/<id>) and profiles (/<user>) whose page carries the site's album markup (the album title heading and media group, or a profile's album links)
fapello	model pages and their posts (/<model>/, /<model>/<n>/) whose page links the model's own media under /content/…/<model>/
fediverse	fediverse accounts, communities and posts (/@name, /users/<name>, /u/<name>, /c/<name>, /notes/<id>, /notice/<id>, /post/<n>, /p/<user>/<id>) on any server whose nodeinfo names its software
feed	RSS and Atom feeds, at any URL: a response served as a feed type, or whose document element is <rss> or <feed>
filejump	public shares (/s/<id>) whose page offers the share's own /s/<id>/download button
filester	file shares (/d/<slug>) loading file_dl.js beside a #downloadButton, and folders (/f/<id>) rendering their #filesGrid or #subfoldersGrid, listing file cards or asking for their password
foolfuuka	FoolFuuka archive threads, posts, searches and board pages (/<board>/thread/<n>/, /<board>/post/<n>/, /<board>/search/…, /<board>/) whose page names the software as its generator and loads its /foolfuuka/ assets, or whose origin's /_/api/chan/ answers the route with archived posts
francetv	video pages (/…/<id>-<slug>.html) and programme listings of France TV whose flight data names video_factory_ids; other .html pages of its app, such as a live channel's, are handed to yt-dlp
gofile	share links (/d/<code>, and the older /<code> and /?c=<code>) whose page is the Gofile app shell, loading its signing script
goonbox	GoonBox image viewers (/img/<id>, also /img/<slug>.<id>) served by its web app; original files are read from the public image metadata API
hls	bare HLS manifests: a response served as a playlist type or opening #EXTM3U, at any URL, or a .m3u8 link; .mpd links are taken in order to refuse them with the reason
imagepond	viewer pages (/i/<code>) naming their stored file (/media/…, /i/<code>/direct) in the player or og:video/og:image, and albums (/a/<code>) of cards linking to viewer pages, on pages named ImagePond
imgur	posts and albums (/a/<id>, /gallery/<slug>-<id>, /<id>) whose page carries window.postDataJSON or its own media as og:image, and account, tag and subreddit listings (/user/<name>, /t/<tag>, /r/<name>) whose listing JSON answers
keep2share	file links (/file/<id>) whose page is the Keep2Share or FileBoom web app shell (its entry-index bundle and chunk-client preload on #root), named Keep2Share or FileBoom; the name chooses the service's own API and proxy group
kemono	Kemono and Coomer creators and posts (/<service>/user/<id>, …/post/<id>) served by the sites' app shell, or whose origin's API returns the creator's matching profile
knuthing	watch pages (/watch/<video>.html) rendered by the KNUTHING theme, with or without their player, and its video listings (/videos/<section>) by their video list
kvs	Kernel Video Sharing pages — videos, member profiles, searches and listings — carrying the licensed player's configuration or script, the platform's generator with playable plain-media flashvars, or its list_videos blocks (asynchronous controls, or a container and its _items list), and video pages whose player is gone by the platform's generator, a block's _items list or its screenshot store
links	a links:<url> prefix — every supported link on the page
lulustream	videos (/<id>, /e/<id>, /d/<id>) whose page names LuluStream and carries its player's HLS source
mediafire	file pages (/file/<key>/…) carrying the #downloadButton link, and folder shares (/folder/<key>) whose page is the MediaFire folder viewer
mediawiki	MediaWiki categories, articles, file pages and upload lists: pages naming a title (/wiki/<title>, index.php?title=) whose generator is MediaWiki and which carry its ResourceLoader configuration, or whose origin's api.php validates an explicit title or generator-marked page as MediaWiki
mega	file and folder links (/file/<handle>#<key>, /folder/<handle>#<key>, /embed/…, #!…!…) whose page is the MEGA web client's boot shell (secureboot.js, #fmholder)
mixcloud	shows (/<user>/<show>/, the trailing slash optional) whose page, named Mixcloud, carries the app's cloudcastQuery for the show
mixdrop	players (/e/<code>) configuring the MixDrop player object (MDCore), and landing pages (/f/<code>) named MixDrop that frame or configure one
niconico	watch pages (/watch/<id>) describing their video in a server-response meta element, and mylists, series and users' uploads rendered by the niconico web client's bundle; live watch pages describing their program in the live client's embedded data, and channels rendered by the channel front end under its portal's name, are handed to yt-dlp
noodlemagazine	watch pages (/watch/<id>) named Noodlemagazine carrying the player's window.playlist or the site's watch script, and search listings (/video/<query>) of its video cards or its "nothing found" notice, rendered with its search script
npo	episodes and series (/start/afspelen/<slug>, /start/serie/<slug>) of NPO Start, served by its app
npr	NPR pages carrying audio players (data-audio): stories, programmes, podcasts and sections
nrk	programmes and series (/program/<id>, /serie/<slug>, /se?v=) of NRK TV, whose page data names programmes by prfId
odysee	claims (/@<channel>/<claim>) whose page, named Odysee, offers the LBRY embed player (/$/embed/), and channels (/@<channel>) and playlists (/$/playlist/<id>) served by its app bundle
ok.ru	videos (/video/<id>, /videoembed/<id>, /live/<id>, a group's or a profile's) whose page mounts the OK video player (OKVideo) with its data-options configuration; group, profile and legacy /dk?st.mvId=<id> links by the embed they name
orf	episodes and series (/video/<id>, /sendereihe/<id>, /profile/…) of ORF ON, whose page carries the runtime configuration (__NUXT__) the API credentials are read from; radio broadcasts, podcasts and stories (/radio/…, /player/<date>/<id>, /podcast/…, /stories/<id>) named for the broadcaster are handed to yt-dlp
pbs	videos (/video/<slug>/), shows (/show/<slug>/) and player links of PBS, by the player a video page embeds or the videos a show page lists
peertube	PeerTube videos, playlists, channels and accounts (/w/<id>, /videos/watch/<id>, /w/p/<id>, /c/<handle>, /a/<handle>) whose page is the PeerTube web client with its inline server configuration, or whose origin's /api/v1/config reports a server version
pixeldrain	pixeldrain files, lists and shared directories (/u/<id>, /l/<id>, /d/<id>, /api/file/<id>) served by the software's viewer, which is handed its API root inline, or, for the software's eight-character ids, or any id whose page was withheld, whose origin's API describes exactly that file, list or directory
pixhost	galleries (/gallery/<code>) of /thumbs/<group>/<file> thumbnails, and viewer pages (/show/<group>/<file>) showing an img#image from /images/
pornhub	video pages (/view_video.php?viewkey=<key>) named Pornhub whose player configuration (flashvars_<n> with its mediaDefinitions) is in the page, and embeds (/embed/<key>) whose key's video page is
pornone	video pages (/<category>/<slug>/<id>/) named PornOne whose player lists renditions as <source> elements named by their frame size
pornpics	galleries (/galleries/<slug>) and category, tag, pornstar and channel listings whose page sets the site's PP_LANG global and marks its links rel-link
radio-canada	pages of ICI Radio-Canada and OHdio whose data layer (IdsCollection) names their media, media console links read from the link alone, and /media-<id> links named for the site or whose page could not be read
raiplay	videos (/video/….html) and programmes (/programmi/<name>) of RaiPlay, whose page mounts its rai-player
rajce	albums (/<user>/album/<slug>) whose page assigns the album's settings, media list included, in a script
redgifs	clips (/watch/<id>, /ifr/<id>) and users (/users/<name>) served by the RedGIFs app, a clip's page naming its embed player
rtbf	media, programmes and live pages of RTBF Auvio (/media/<slug>-<id>, /emission/<slug>-<id>) whose page addresses the Auvio API, and RTBF pages embedding Auvio players
rtpplay	episodes and seasons (/play/p<id>[/e<id>]) of RTP Play, whose page carries its player or its episode list
rtve	videos and programmes (/play/videos/…, /v/<id>) of RTVE Play, whose page names its asset or programme id
rumble	videos (/v<id>-<slug>.html, /embed/v<id>) that the origin's Rumble oEmbed endpoint describes, and channels, users and playlists (/c/<name>, /user/<name>, /playlists/<id>) served by its app bundle
ruv	programmes and episodes (/<station>/spila/<slug>/<id>) of RÚV's player, whose page configures its GraphQL endpoint (window._GQL_)
sendvid	videos (/<id>, /embed/<id>, and /<id>.mp4, which redirects to the page) whose page, named Sendvid, carries the player's video_source
soundcloud	tracks and sets (/<artist>/<track>, /<artist>/sets/<set>) whose page is the SoundCloud web client hydrated with the sound (__sc_hydration) and offering its widget player, private tracks and sets by their secret token (…/s-<token>) hydrated the same way, and artists (/<artist> and its tabs) hydrated the same way
srf	programmes and shows of Play SRF (a urn= or a show's id= in the query) whose page names its urn:srf: media
streamable	videos (/<id>, /e/<id>, /s/<id>/<slug>) whose page, named Streamable, offers the site's player (/t/<id>) for embedding, or whose video the player's document describes
streamtape	videos (/v/<id>, /e/<id>) whose page assembles the Streamtape robotlink media link
suvobox	albums (/a/<id>) listing thumb-item tiles, and file pages (/f/<id>) whose download button is the file's own /f/<id>?dl=1 link
svtplay	videos (/video/<id>) and programme pages of SVT Play whose page data names the player API's videoSvtIds
tnaflix	video pages (/<category>/<slug>/video<id>) named TNAFlix whose player lists its renditions as <source> elements
turbo	video pages (/embed/<id>, /d/<id>, /v/<id>) whose player or download page calls the /api/sign?v= signing endpoint, or that frame that player
vidmoly	watch pages (/w/<code>, /v/<code>) of the VidMoly app naming the file's player, and players (/embed-<code>.html) whose JWPlayer source list names VidMoly; aliases, and a bare /<code> an alias turned away, are validated against the canonical player
vids.st	videos (/v/<id>, /e/<id>) whose page carries the player's playerConfig with its videoUrl
vimeo	player pages (/video/<id>) carrying window.playerConfig, and watch pages (/<id>, /<id>/<hash>, channel, group, album and showcase videos, review and manage links) of a site named Vimeo that offer that player for embedding, or whose video's player page does
voyeurking	video pages (/video/<slug>) and category and collection listings (/categories/<name>, /collection/<slug>) whose React Router data names the routes/video.$slug, routes/categories or routes/collection route
vrtmax	programmes and episodes (/vrtmax/a-z/…) of VRT MAX, served by its app
wetransfer	download pages (/downloads/<id>/<hash>) and the short links (/t-<code>) that redirect to them, served by the WeTransfer app with the transfer's securityHash in its page data
xhamster	video pages (/videos/<slug>-<id>) named xHamster that carry the site's page state (window.initials) and player settings
yandex	video search previews (/video/preview/<id>, /video/search?filmId=<id>, their touch forms) built by the Yandex search front end, naming the source video in its viewer
yle	programmes and series (…/1-<id>) of Yle Areena, whose page carries the Areena web app's configuration
youtube	videos (/watch?v=<id>, /shorts/<id>, /embed/<id>, /clip/<id>), playlists, channels and their tabs, searches, hashtags and music pages whose page carries the YouTube web client's configuration (ytcfg) beside its player or page data or under the YouTube Music name, or the consent form that stands in for such a page; a video, embed, playlist or /channel/UC<id> link whose probe was refused or failed
zdf	videos and show pages (/…/<slug>-<number>) of the ZDF Mediathek, whose page data names its PTMD player templates, and documents of its metadata API (/mediathekV2/document/<id>)

Features

See the feature guide for transfer behavior, queue controls, naming, image-board support and recovery details.

Screenshots

Light theme, mobile layout and terminal mode

Use -debug=false for animated terminal progress. Debug logs are enabled by default; pipes and log files always receive plain progress lines.

Documentation

Start with the documentation index. If you are reading this inside a release archive, the full online documentation is available on GitHub.

Configuration

Flags and environment variables cover download locations, concurrency, bandwidth, queue state and proxy settings. Flags take precedence over environment variables.

Optional helpers

Media helpers, local OCR and proxies explains which workflows need extra tools, how to install them, and their platform limits. The Docker image includes them.

HTTP API

Use the HTTP API reference to submit downloads, change settings or consume the event stream.

Network access and trust

HeapLeach is a single-user service without authentication. Anyone who can reach its API can inspect the queue, submit downloads and change the destination. Use a loopback address locally; remote access needs an authenticated HTTPS reverse proxy and restricted access to the backend port. Read the network guide.

Architecture

The architecture guide follows a URL from extraction to disk and explains the package dependency rules.

Development

To try a source build on Linux, install Git, GNU Make and Docker, then:

git clone https://github.com/JohanLindvall/HeapLeach.git
cd HeapLeach
make build
./bin/heapleach

Docker produces a Linux executable. On macOS or Windows, use the prebuilt release, run the Docker image, or follow the native development setup.

For hot reload, install the Go toolchain and Node.js 24, run npm ci in frontend/, then make dev from the root. See CONTRIBUTING.md for the complete setup and checks, and the development reference for build targets and releases.

Jump to source layout, adding a host or release packaging.

Contributing

Bug reports, documentation improvements, UI fixes and new extractors are welcome. You do not need to know the whole codebase to help.

• Read the contribution guide for setup and suggested starting points.
• Report a bug or request a feature.
• Get help or report a vulnerability privately.
• Follow the code of conduct.

If HeapLeach is useful to you, star the repository to help others find it. Sharing a clear bug report or a small improvement helps the project too.

Star history

About the name

Why “HeapLeach”?

Heap leaching is a mining method: crushed ore is piled into a heap and irrigated from above, and the solution percolates down through it, dissolving out the metal as it goes and draining to a pad at the bottom. This does the same to a heap of links. The extractors percolate through whatever the pages are made of — players, listings, signed redirects, encrypted payloads — and what is worth keeping drains out into a folder.

Notes

The operating notes explain queue checkpoints, partial files and live-site tests. Use modest concurrency and respect host limits.

License

HeapLeach is MIT licensed. Bundled OCR assets and external media helpers retain their own licenses; see the notices in ocr/ and the OCR packaging guide.