timing/wrap: runs 100% locally · translation: optional local LLM (WebGPU)
Subtitle translator & SRT fixer
Fix drifted timing, wrap long lines to a character limit, and optionally translate — using a language model that runs on your own GPU, not a per-minute cloud API.
Your .srt file is parsed and edited in memory in this tab. If you use translation, the model itself is downloaded once from a public model hub straight to your browser — your subtitle text is never sent to us.
Drop a .srt file here, or click to browse
Standard SubRip subtitle format.
No file yet — drop a .srt file above to get started.