tylerbcrawford/subgeneratorr v2.4.0-beta.2

v2.4.0-beta.2 — LLM subtitle translation (beta, UI cleanup)

on GitHub

Updated beta for the Translate Subtitles feature. Supersedes v2.4.0-beta.1 — same translation feature, cleaner UI. Opt-in: this publishes its own image tag and does not change your :latest.

Changed since v2.4.0-beta.1

• Target languages are now chosen from an "Add a language" dropdown with removable chips (replacing a long wall of checkboxes), styled to match the transcription language selector.
• Layout/mobile fixes: the Translate section no longer sits between the "Transcription Settings" toggle and its panel (settings no longer appear to vanish), the dropdown is centered, the description blurb was trimmed, and the provider-status badge is compact.

What it does

Translate a generated SRT into one or more languages with an LLM, preserving timing exactly and writing language-tagged sidecars (.spa.srt, .fre.srt, …) that Plex, Jellyfin, and Emby auto-detect. Providers: Anthropic Claude, OpenAI GPT, Google Gemini, or a local Ollama model (free, fully offline; cost shows $0.00 — use a 3B+ model such as qwen2.5:7b).

Try it

Point your web and worker services at the new tag (comment out their build: blocks while testing the image):

  web:
    image: ghcr.io/tylerbcrawford/subgeneratorr-web:2.4.0-beta.2
  worker:
    image: ghcr.io/tylerbcrawford/subgeneratorr-worker:2.4.0-beta.2

docker compose pull
docker compose up -d

Open the Web UI, select a video that already has a subtitle, open Translate Subtitles, add your languages, pick a provider, and Translate.

Ollama in Docker: set OLLAMA_HOST=http://host.docker.internal:11434 in .env (the example compose maps host-gateway).

Please report results in the pinned Beta feedback issue #3 (provider/model, languages, your setup, and any docker compose logs worker output if something fails). Thanks for testing!