This skill is for the narrow but very common job of taking an existing subtitle file that no longer matches the final edit and bringing it back into sync with the finished audio. An agent using Subaligner can ingest a video or audio file plus a subtitle track, choose the appropriate alignment mode, run the retiming pass, and return a corrected subtitle file in a standard format such as SRT or WebVTT. It can also handle batch alignment when a team needs to fix many files in one pass.

Invoke this when the user already has subtitles or script text, but timing drift has crept in because of editorial changes, frame rate differences, distribution edits, or language-localized versions. This is where using the product normally by hand becomes tedious: the agent can pick the alignment mode, run the job, and keep the process repeatable across many assets.

The scope boundary matters. This is not a general video editor, not a streaming platform, and not a broad speech-to-text suite entry. The skill is specifically about subtitle timing recovery and related alignment workflows. It may use Subaligner transcription or translation features when they directly support the alignment task, but the core value is recovering accurate subtitle timing against a finished media file.

Integration points are concrete: FFmpeg is required upstream, Docker is available for containerized runs, and the tool can fit into localization pipelines, post-production QA, transcription cleanup, and archive remediation workflows.