I built an open-source tool that turns any NAS into a fully automated subtitle-generation pipeline — running completely locally, no cloud services, no data leaving your network.
If you keep movies/TV shows on a Synology/QNAP/TrueNAS box, this project makes the NAS generate subtitles automatically using on-device models.
Highlights
Local speech-to-text with Whisper-compatible engines
No external API calls — privacy by design
Simple plug-and-play: drop media in a folder → subtitles appear
Multi-thread + GPU acceleration support
Auto-language detection, timestamp correction, subtitle formatting
Web UI for queue monitoring & batch processing
Works on Synology, QNAP, Unraid, TrueNAS, Docker everywhere
Why I built this
Existing subtitle tools either rely on cloud APIs (privacy concern), require manual processing, or don’t integrate well with NAS workflows.
So this project aims to provide:
A fully autonomous subtitle engine for home servers.
Repo + Docs:
https://subtitlesdog.com/en/nas-subtitler
https://github.com/subtitlesdog/nas-subtitler
Happy to hear feedback from HN folks — feature suggestions, performance improvements, or architecture critiques are more than welcome.
If you keep movies/TV shows on a Synology/QNAP/TrueNAS box, this project makes the NAS generate subtitles automatically using on-device models.
Highlights Local speech-to-text with Whisper-compatible engines No external API calls — privacy by design Simple plug-and-play: drop media in a folder → subtitles appear Multi-thread + GPU acceleration support Auto-language detection, timestamp correction, subtitle formatting Web UI for queue monitoring & batch processing Works on Synology, QNAP, Unraid, TrueNAS, Docker everywhere
Why I built this
Existing subtitle tools either rely on cloud APIs (privacy concern), require manual processing, or don’t integrate well with NAS workflows. So this project aims to provide:
A fully autonomous subtitle engine for home servers.
Repo + Docs: https://subtitlesdog.com/en/nas-subtitler
https://github.com/subtitlesdog/nas-subtitler
Happy to hear feedback from HN folks — feature suggestions, performance improvements, or architecture critiques are more than welcome.