Reference documentation
The pages you'll keep coming back to: installation on every platform, the full flag reference, authentication details and the deep internals.
Getting started
New here? Start with these. Install the right build for your OS, learn the flags every command uses, and sort out how login works before you need it. Get these three down and the rest of the docs become reference you dip into, not reading you have to do.
Installation
Windows, Linux, macOS, ARM, the framework build, Docker and headless boxes — every way to get DepotDownloader running, including the ones I'd skip and why.
read the docs → referenceCLI reference
Every flag verified against the source: targets, auth, filters, branches, behaviour — with hidden aliases, conflicts and copy-paste recipes. Filterable.
open the reference → accountsAuthentication
Anonymous, password, QR, Steam Guard — and exactly what gets stored on disk when you say -remember-password, plus how to revoke it all.
read the docs →The Steam content model
The concepts that make everything else click. Once you can picture how an app splits into depots, how a manifest pins one version, and where to read those IDs on SteamDB, the commands stop feeling like magic incantations and start feeling like addresses.
Manifests & request codes
The deep dive: what's inside a manifest, how chunks are encrypted and deduplicated, and the exact server-side check that decides whether your download lives or dies.
read the deep dive → reference · how-toFinding your IDs on SteamDB
The lookup every command needs: how to find the App, Depot and Manifest ID on SteamDB, how to pick the right depot, and the Build ID vs Manifest ID trap.
read the how-to → encyclopediaThe warehouse dictionary
23 terms in plain language — app, depot, manifest vs build ID, branch, chunk, depot key, request code — each with the deep dive one click away.
read the glossary → referenceBranches & betas
Pull an old version by name with -branch instead of hunting a manifest — and the crucial difference between a branch (a moving pointer) and -manifest (an exact pin). Plus password branches and the silent public fallback.
read the docs →Surgical downloads & migration
Sometimes you don't want the whole game. These cover pulling a single file out of a 60 GB depot, the old Steam console trick and why it broke, and where DepotDownloader picks up the jobs SteamCMD simply can't.
Filelists & -manifest-only
Pull one file out of a 60 GB depot: filelist syntax, regex patterns, and the diff workflow that shows exactly what a patch changed.
read the docs → reference · migrationThe download_depot command
Steam's hidden console trick for old versions — how it worked, why it now fails or only grabs the current build, and the one-line migration to DepotDownloader.
read the how-to → reference · comparisonSteamCMD can't get an old version
SteamCMD only installs the current build — it can't pin an old one. Why, the DepotDownloader bridge, and how to roll a dedicated server back by manifest.
read the how-to →Automation & self-hosting
Running it unattended — in CI, on a NAS, behind a LAN cache. The recipes here are the ones I actually keep around: a session that survives a reboot, a container that behaves on appliance OSes, and a cache that turns twenty downloads into one.
Automation: CI, Docker, cron
Running it unattended — seeding sessions without prompts, the Docker recipe, a GitHub Actions example, exit-code handling and CDN politeness.
read the docs → self-hostOn a NAS & in Docker
Run it on Synology, unRAID, QNAP or TrueNAS: the container image, the volume mounts and PUID/PGID permissions that matter, persistent login, and per-appliance notes.
read the how-to → self-host · networkDownloading through a Lancache
What -use-lancache really does — auto-detection, the HTTP downgrade, max-downloads 25 — when caching helps, and the honest split between DepotDownloader (keep a copy) and SteamPrefill (warm the cache, keep nothing).
read the docs →