Documentation
Everything you need to install, run and update Smart Transcode. Start with the requirements, read the page on user and group IDs, then follow the install guide for your setup.
Choose how to run it
🖥Linux service
A systemd service on the machine itself. One command, no questions. Needs Python and an ffmpeg with your encoder.
🐳Docker
One image with every encoder inside. Compose files for NVIDIA, Intel, AMD and CPU.
🦭Podman
The same image, rootless or rootful, with GPU access, SELinux notes and a Quadlet unit for boot.
Not sure? Use Docker or Podman: the image brings an ffmpeg with every encoder the tool uses, so GPU encoding works without building anything.
Before you install: user and group IDs
Smart Transcode replaces your files, and Linux allows that only to the right numeric user and group. In Docker these are PUID and PGID.
Getting them wrong gives empty libraries or "permission denied" at the end of every conversion. Read this page first.
All pages
| Page | What it covers |
|---|---|
| Requirements | System, software and GPU needs for each way of running it. |
| User and group IDs | How file ownership works, PUID / PGID / UMASK, shared groups, network shares and a symptom checklist. |
| Linux service | Install with install.sh, its options, and GPU notes for bare metal. |
| Docker | The setup helper and compose files for NVIDIA, Intel, AMD and CPU. |
| Podman | Rootless and rootful, GPUs, SELinux, starting at boot with Quadlet. |
| First steps | Add folders, pick the encoder, convert a first file. |
| Upgrading | Upgrade with the latest package or image, roll back, move from the service to a container. |
| Reverse proxy | HTTPS in front of the page with nginx, Traefik, Caddy or Apache. |
| Troubleshooting | Unreachable page, missing GPU, permission errors, passwords, where the data is. |
The package you download also contains these guides, the compose files and the reverse-proxy samples. Download