How is subtitle diffing different from standard text or Git diffing?

Standard Git diff tools compare raw line numbers. When a single cue is inserted or deleted, every subsequent line shifts, causing line-based diff tools to report the entire rest of the file as changed. SubPreflight subtitle diffing aligns cues by millisecond timecodes and dialogue overlap, isolating exact text edits and retimed shifts.

Can I compare subtitles in different file formats, such as SRT vs WebVTT?

Yes. You can compare an SRT file against a WebVTT (.vtt), SubViewer (.sbv), or SubStation Alpha (.ass) file. The tool parses each format into standard timestamped cues before comparing dialogue and timing.

How are retimed cues detected versus edited text?

If a cue contains the exact same spoken dialogue but starts or ends at a different timestamp, it is flagged as Retimed with the exact millisecond offset (for example, +500ms). If the timestamps overlap but words were modified, it highlights word-level additions and deletions.

Can I export a change summary or audit log for clients and QC teams?

Yes. You can export a CSV audit report containing cue numbers, before and after timecodes, dialogue differences, and time shifts, or download a text summary and clean modified subtitle file.

Are my subtitle files or video scripts uploaded to a remote server?

No. All subtitle parsing, diffing, and text analysis take place entirely in your web browser using local JavaScript memory. Your confidential video transcripts and project drafts never leave your device.

.SRT.VTT.SBV.ASS

Subtitle Diff & Version Compare

Compare two subtitle revisions side by side. Track added, deleted, modified text, and retimed cues with millisecond accuracy. 100% private in-browser.

100% In-Browser · Zero Upload

Zero latency · Client-side Web Worker

Revision 1 (Original)

Drop original file or browse

Accepts SRT, VTT, SBV, ASS

Revision 2 (Modified)

Drop revised file or browse

Accepts SRT, VTT, SBV, ASS

Tip: You can drop two files together into either card to populate both revisions.

Why Subtitle Diffing Requires Cue-Aware Parsing

Comparing subtitle files with conventional diff tools like Git or raw text comparators frequently fails. If a single caption is added at minute 2:00, every downstream cue number in a SubRip (.srt) file increases by one. A line-based diff marks every subsequent line as deleted and re-inserted, creating hundreds of false positives.

SubPreflight solves this by parsing subtitles into structured cue models. It matches cues based on timestamp proximity and dialogue similarity, allowing editors, localization managers, and quality control teams to see:

Word-level corrections: Exact edits to character names, spelling, and phrasing.
Retimed cues: Dialogue shifts caused by video re-cuts or commercial insertion, measured in exact milliseconds.
Added captions: Brand-new dialogue lines or translator additions.
Deleted captions: Cut lines or removed scene dialogue.

Workflow for Client & Translation Approvals

When delivering captions to broadcast networks, streaming platforms, or corporate clients, revision tracking is essential:

Load your master caption file into Revision 1.
Load the client or editor revised caption file into Revision 2.
Filter by Differences Only to inspect every modification.
Export a CSV Audit Report to provide documented proof of all approved script changes.

Frequently Asked Questions

How is subtitle diffing different from standard text or Git diffing?: Standard Git diff tools compare raw line numbers. When a single cue is inserted or deleted, every subsequent line shifts, causing line-based diff tools to report the entire rest of the file as changed. SubPreflight subtitle diffing aligns cues by millisecond timecodes and dialogue overlap, isolating exact text edits and retimed shifts.
Can I compare subtitles in different file formats, such as SRT vs WebVTT?: Yes. You can compare an SRT file against a WebVTT (.vtt), SubViewer (.sbv), or SubStation Alpha (.ass) file. The tool parses each format into standard timestamped cues before comparing dialogue and timing.
How are retimed cues detected versus edited text?: If a cue contains the exact same spoken dialogue but starts or ends at a different timestamp, it is flagged as Retimed with the exact millisecond offset (for example, +500ms). If the timestamps overlap but words were modified, it highlights word-level additions and deletions.
Can I export a change summary or audit log for clients and QC teams?: Yes. You can export a CSV audit report containing cue numbers, before and after timecodes, dialogue differences, and time shifts, or download a text summary and clean modified subtitle file.
Are my subtitle files or video scripts uploaded to a remote server?: No. All subtitle parsing, diffing, and text analysis take place entirely in your web browser using local JavaScript memory. Your confidential video transcripts and project drafts never leave your device.