Subtitles for everything you watch.Found, timed and translated, automatically.

What it does

If you have never automated subtitles before, this is the whole job. Bazarr+ does each of these on a schedule, for everything in your library, and tells you when it could not.

No media stack required.

Bazarr+ used to be able to work only on what Sonarr and Radarr told it about. Since v2.7.0 it opens on Discover, and Discover works on its own.

If you run Seerr, Jellyseerr or Overseerr, a title you do not have yet can be requested from its page in Discover. The request goes in as the Seerr owner, so it is approved and dispatched immediately rather than queued for somebody to look at. Requesting titles from Discover.

What it connects to

Seven applications, in two groups. One group tells Bazarr+ what media exists. The other is told when a subtitle lands. They are configured differently and they fail differently, which is worth knowing before you add one.

Two things worth knowing before you configure them. Emby and Silo find a file by its path, so they need path mappings; Jellyfin and Plex match another way and have never asked for them. And Plex is on the shared layer for refreshes only: its account panel, server discovery and webhook helpers still read the single Plex connection in settings. How media server connections work.

The interface

Screenshots from the running application. Nothing here is a mockup. v2.7.0 rebuilt what you meet first: Discover is the homepage, the setup wizard asks what you want to do instead of insisting on Sonarr, and Statistics has a page of its own under System.

What this fork adds

Bazarr+ is a hard fork of Bazarr by morpheus65535, and it keeps everything that already worked. This is what it puts on top. If you are new here, you can skip this table entirely.

Install

Start with standalone subtitle search, or connect Sonarr, Radarr or Sportarr to automate your library. Media folders are optional.

Quick install

Run this on your server. The installer creates the Docker Compose stack and starts Bazarr+. Leave the movie, TV and sports paths empty to use Discover on its own. Review the script.

curl -fsSL https://lavx.github.io/bazarr/install.sh | bash

When it finishes, open the address printed by the installer. The installer pre-enables three providers, so the app skips the first-run wizard and opens straight in Discover; to walk the wizard anyway, use Settings > General > Run first-time setup. If you browse from another machine, use your server's IP or hostname instead of localhost.

Prefer Docker Compose?

With Docker and Docker Compose installed, save this as compose.yaml. It works as-is for standalone Discover. For a library, uncomment only the folders you use and replace each /path/to/... with its location on your server. Set PUID and PGID to the user and group that own those folders.

services:
  bazarr:
    image: ghcr.io/lavx/bazarr:latest
    container_name: bazarr
    restart: unless-stopped
    ports:
      - 6767:6767
    environment:
      - PUID=1000
      - PGID=1000
      - TZ=UTC
    volumes:
      - ./config:/config
      # Optional: uncomment only the media folders you use.
      # - /path/to/movies:/movies
      # - /path/to/tv:/tv
      # - /path/to/sports:/sports
    read_only: true
    tmpfs:
      - /tmp:size=512m
    cap_drop:
      - ALL
    cap_add:
      - CHOWN
      - SETUID
      - SETGID
    security_opt:
      - no-new-privileges:true

From the folder containing compose.yaml, run:

docker compose up -d

Open http://localhost:6767, or use your server's IP or hostname instead of localhost. Choose standalone Discover in the setup wizard, or connect your library under Settings > Connections. The TV, movie and sports mounts are for Sonarr, Radarr and Sportarr libraries respectively; only /config is needed for standalone search.

Already using Bazarr? Stop the old container, back up its /config directory, then change its image to ghcr.io/lavx/bazarr:latest while keeping your existing mounts. See the migration guide.

Optional extras: add FlareSolverr for OpenSubtitles.org Cloudflare challenges or the AI translator sidecar for translation. The getting started guide covers both.

Questions

Roadmap

Where Bazarr+ is going, nearest first. There are no dates here, because a self-hosted project ships when the work is right rather than when a quarter ends. Plans shift; the direction has not.

1. Shipped

   The v2 line so far

   v2.7.2 is a stability and security patch on the Atlas line. It masks credentials in logs and diagnostics, guards the request host while login is off, moves the editor's native playback to a signed token, and adds library filters for subtitle completeness, tags and score. v2.7.1 is a stability and security patch on the Atlas line. It moves API keys out of request URLs, runs post-processing commands as argument lists, and binds Provider Hub trust to the catalog source. v2.7.0 “Atlas” opens on Discover: search and pull subtitles for any title, whether or not it is in your library, and run with no library at all. It adds Sports as a third media type, folds Emby, Silo, Jellyfin and Plex onto one multi-instance layer, requests titles in Seerr, and gives Statistics a real page. Before it, v2.6.2 “Clockwork” keeps subtitle history and sync outputs with the right media, saves useful partial OpenRouter translations with an incomplete status, lists uploads before background sync finishes, and adds explicit OpenRouter routing choices. Searches continue past unusable archives from catalog providers. The v2.6 line also brought whole-season subtitle downloads, stricter sync validation, and adoption of upstream Bazarr databases. Before it, v2.5 “Murmuration” brought many Sonarr and Radarr instances managed as one, per-instance subtitle settings and archive uploads.

   Every release, with full notes

2. Next

   Coming on the v2 line
   • A rebuilt translator. Translation moves in-process, with a catalog of models, configuration per engine, and a choice of engine per language profile, instead of a separate sidecar container.
   • Search inside your subtitles. Find the episode from a line of dialogue you half remember, across everything you have.
   • Install it like an app. Native installers for Windows, macOS and Linux, each one carrying its own Python and a tray icon, so Bazarr+ runs on a desktop with no Docker and no terminal.

3. v3.0.0

   Several v2 releases away

   Bazarr+ instances sharing subtitles directly with each other, peer to peer, with no central server in the middle and nothing shared that you did not choose to share. It is the largest thing on this list by a distance, which is why it carries the major version and why a good deal of the v2 line comes first.