ArgonFetch

Appearance

What is ArgonFetch? ​

ArgonFetch is a self-hosted web application that turns a media link into a file you keep.

Paste a URL, and ArgonFetch resolves it, lists every quality the source actually offers, and serves the one you pick back as an ordinary browser download. There is nothing to sign up for and nothing to configure - not even Spotify credentials.

  • No API keys. Spotify metadata is read from the public track page, so there is no developer app to register and no secret to rotate.
  • Audio or video. Renditions come straight from the source, with a size and bitrate on each so you can see what you are choosing.
  • Web interface and REST API. The UI is one input box; the API is a handful of GET endpoints, documented at /swagger and checked in as openapi.json.
  • One container plus a database. Postgres holds the resolved-media cache. Everything else runs in the app container.

Pre-release

ArgonFetch is under development and the main branch may not behave as documented. For a working build, use a tagged image or the releases page.

How a download works ​

  1. You paste a URL. The frontend calls GET /api/Fetch/GetResource?url=....
  2. ArgonFetch resolves it through yt-dlp and returns the title, author, cover and the list of available renditions - each with a key.
  3. You pick one. The browser opens GET /api/Stream/Media/{key} (or /api/Stream/Combined/{key} when video and audio have to be muxed) and the bytes come back as a file download.

Nothing is stored on disk for you to clean up afterwards; the stream is proxied through as it is fetched.

Formats ​

Audio is delivered in whatever container the source uses - Opus in WebM for YouTube - because passing the original bytes through is both faster and better than re-encoding them to MP3. Every response states its mimeType. MP3 is still available on request: add ?format=mp3 to the stream URL, which is also what happens automatically for sources in a container ArgonFetch does not recognise. Video is delivered as MP4.

Media tooling ​

yt-dlp and FFmpeg are not built into the image. The container fetches both when it starts and reports itself as under maintenance until they are ready - the web UI shows a maintenance screen and fetches answer 503 for those few seconds. yt-dlp then updates itself every 12 hours, so extractor fixes land without rebuilding anything, and a restart is enough to recover from a broken version.

Next ​

Released under the GPL-3.0 License.

ArgonFetch