silo-server is a free, open source digital asset management (dam) project written in Go and released under AGPL-3.0. It has 1,043 GitHub stars, 114 forks and 353 open issues, and was last pushed 11 hours ago. On this registry it ranks #20 of 27 tracked projects in Digital Asset Management (DAM), with 5 head-to-head comparisons available.

What is silo-server?

Silo is a self-hosted media streaming server, written in Go with a React web app and deployed with Docker, for people who want to run their own film, series, audiobook, ebook, podcast and manga library.

What it is

Silo keeps one catalog for films, series, audiobooks, ebooks, podcasts and manga, matches items through metadata plugins such as TMDB and TVDB, and plays them back with direct play when possible, remux when needed and transcoding otherwise, including VA-API, Quick Sync and NVENC hardware acceleration. Libraries, users, providers, storage, search and playback are configured in an admin interface rather than in configuration files. It is written in Go, licensed AGPL-3.0 and sits in this registry's Content & Publishing / Digital Asset Management category.

The project lives in the self-hosted media server ecosystem and replaces the incumbent server rather than the clients: it exposes a Jellyfin/Emby-compatible API, so applications such as VidHub, Findroid and Infuse can connect to it, while it also ships its own web app. That lets an operator consolidate catalog, metadata matching, playback and household management into one deployment instead of stitching several services together.

Key capabilities

  • Direct play, remux or transcode per item, with VA-API, Quick Sync and NVENC hardware acceleration.
  • One catalog for films, series, audiobooks, ebooks, podcasts and manga, matched through metadata plugins such as TMDB and TVDB.
  • A Jellyfin/Emby-compatible API alongside the bundled web app, used by clients such as VidHub, Findroid and Infuse; client coverage varies.
  • Household profiles with their own watch state, library access and parental controls.
  • Admin-interface configuration of libraries, users, providers, storage, search and playback, with no config files.
  • Transcode nodes that need only the server's URL and a key, and proxy nodes that share its PostgreSQL and Redis.
  • An omnibus Docker image bundling PostgreSQL with pgvector and Redis, plus Meilisearch when it is turned on.

Who uses it and how

  • Households running a single integrated server and using profiles for per-member watch state, library access and parental controls.
  • People who already have Jellyfin or Emby clients, pointing VidHub, Findroid or Infuse at Silo through the compatible API.
  • GPU-equipped deployments that offload transcoding to transcode nodes on other hosts, reached on the server's web port with a URL and a key.
  • Multi-service setups where proxy nodes connect to the primary server's PostgreSQL and Redis.
  • Operators on Docker Compose 2.24 or newer who are comfortable editing .env to set MEDIA_ROOT and SECRET_KEY.

Getting started

Clone the repository, copy .env.example to .env, generate a SECRET_KEY and set MEDIA_ROOT to the absolute path of your media, then run docker compose up -d with Docker Compose 2.24 or newer; the default Compose file starts the omnibus image with bundled PostgreSQL and Redis. Hardware transcoding and external PostgreSQL and Redis access are commented out in docker-compose.yml and must be uncommented.

How it compares

Silo sits beside Jellyfin and Emby rather than apart from them: it speaks a Jellyfin/Emby-compatible API so clients written for those servers can connect, while adding its own React web app and admin interface. It also carries beta Audiobookshelf compatibility for audiobooks and ebooks, though that work sits outside the 1.0 support promise. Under AGPL-3.0 it is self-hosted and self-operated, with no hosted option described in the material available here.

When to use it — and when not to

Choose Silo if you want an actively developed server that reuses Jellyfin and Emby clients and can tolerate pre-release churn, since APIs, configuration and database migrations may change before the first stable release and the README advises backing up before updating. Avoid it if you need a stable release or a firm promise for audiobooks and ebooks, which are labeled beta and will be replaced by a consolidated Books effort with no assigned release date, while the 1.0 milestone covers only movies and series. Upgrading a pre-omnibus install also takes care: an install that keeps Silo, PostgreSQL and Redis in separate containers must set COMPOSE_FILE=docker-compose.advanced.yml before updating, and moving into the single container is a separate step.

project readme (upstream, from github) — read inline

Silo

A self-hosted media server for films, series, audiobooks, ebooks, podcasts, and manga.

Latest GitHub release Container image on GHCR Continuous integration Go 1.27 React 19 AGPL-3.0-or-later license

Quick start · Documentation · Builds & releases · Discord · Support Silo · Contributing


[!WARNING] Silo is pre-release. APIs, configuration, and database migrations may change before the first stable release. Back up your deployment before updating.

1.0 release scope

The 1.0 milestone covers movies and series. Audiobooks, ebooks, and Audiobookshelf compatibility keep working as they do today and are labeled beta: they are outside the 1.0 support promise and will be replaced by a consolidated Books effort with no assigned release date. The broader capabilities below include those beta implementations, not a promise that every media type is supported in 1.0. Existing library data is preserved when upgrading.

What Silo does

Play

Direct play when possible, remux when needed, transcode otherwise, with VA-API, Quick Sync, and NVENC hardware acceleration.
Organize

One catalog for films, series, audiobooks, ebooks, podcasts, and manga, matched through metadata plugins such as TMDB and TVDB.
Connect

Use the included web app, or the Jellyfin/Emby-compatible API with clients such as VidHub, Findroid, and Infuse. Client coverage varies.
Share

Household profiles with their own watch state, library access, and parental controls.
Manage

Libraries, users, providers, storage, search, and playback are configured in the admin interface, not in config files.
Scale

Start with one integrated server; add transcode nodes that need only its URL and a key, or proxy nodes that share its PostgreSQL and Redis, when you need to.

Quick start

Requires Docker Compose 2.24 or newer. The default Compose file runs the omnibus image: Silo with a bundled PostgreSQL (with pgvector) and Redis, plus Meilisearch when you turn it on.

git clone https://github.com/Silo-Server/silo-server.git
cd silo-server
cp .env.example .env
chmod 600 .env
key="$(openssl rand -base64 48)" && printf '\nSECRET_KEY=%s\n' "$key" >> .env

Set MEDIA_ROOT in .env to the absolute path of your media, then:

docker compose up -d

Open and complete onboarding.

Hardware transcoding (VA-API or NVIDIA) and PostgreSQL and Redis access for proxy nodes on other hosts are commented out in docker-compose.yml; uncomment the lines you need. Transcode nodes reach the server on its web port.

Updating an install from before the omnibus image

An install that runs Silo, PostgreSQL, and Redis in separate containers keeps that layout when its .env sets COMPOSE_FILE=docker-compose.advanced.yml; set it before updating. docker-compose.yml is an example of a new install and does not look for an earlier one: started on that data folder, it replaces the old silo container and comes up on a new, empty database with setup open, while the earlier PostgreSQL and Redis keep running with their data untouched. Moving into the single container is a separate step. The separate GPU and node Compose files are gone, so a COMPOSE_FILE that names one fails until it names only the deployment file you use, with the matching lines in it uncommented. See Moving from the multi-container stack.

The user manual covers the rest, starting with installation: backing up SECRET_KEY, storage, GPU acceleration, Meilisearch, external PostgreSQL and Redis, transcode nodes, PostgreSQL tuning, backups, and updates.

Builds and releases

Until the first release, default-branch images carry an ordered build-N tag and the full commit SHA alongside latest. Build numbers order published images; they are not release versions. Release versioning defines each tag and the SemVer contract.

Documentation

Community and contributions

Questions and discussion: Discord. Bugs, install problems, and performance issues: the GitHub issue forms, which ask for reproduction steps and raw logs.

Native clients are developed in silo-apple and silo-android. Client-visible API, authentication, playback, or metadata changes should consider both.

Right now, the most helpful way to contribute is a clear, accurate issue. We implement and review changes through our own workflow, and pull request volume has outgrown what we can review carefully, so a good issue usually helps us more than a pull request.

Feature ideas go in Discussions → Ideas, not issues.

Pull requests are still welcome from contributors who've had one merged in a Silo repository, or when a maintainer asks for one; a feature pull request also needs an issue in this repository that a maintainer has labeled accepted. Other pull requests are closed automatically without review; when a maintainer asked for one, they reopen it. CONTRIBUTING.md explains what makes an issue useful.

Supporting Silo

Silo is free and open source, and it will stay that way. It's built in spare time and paid for out of pocket.

Sponsor Silo on GitHub

Sponsorship pays for AI development tooling (Claude, Codex), the push-notification relay, and future project costs. You can help just as much without spending anything:

Sponsors


Code review is supported by Macroscope through its open source program.

License and trademarks

Silo's source code is licensed under the GNU Affero General Public License v3.0 or later (AGPL-3.0-or-later), with an additional permission to combine it with platform SDKs such as the Windows App SDK. See LICENSE and PLATFORM-EXCEPTION.md.

The Silo name, logo, and wordmark are trademarks of Silo Media L.L.C. and are not covered by the AGPL. Forks and redistributions may use the code but must not use the Silo brand as their identity and must remove or replace the brand assets. Publishing a Silo-branded app to an app store requires written permission. See TRADEMARK.md for permitted referential use, including phrases such as "compatible with Silo."

Frequently asked questions

Is silo-server free to use?

silo-server is open source under the AGPL-3.0 licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does silo-server do?

Self-hosted media streaming server with a Go backend, React web UI, Docker deployment, transcoding, and Jellyfin-compatible APIs.

What is silo-server written in?

silo-server is primarily written in Go. Its source is publicly available at https://github.com/Silo-Server/silo-server, and it has 1,043 GitHub stars.