Work with transcriptsFree · No account needed
Search multiple transcripts for a word or phrase
Find the passage you remember across transcripts you supply. Search a literal word or phrase, compare its surrounding context, and carry useful matches into research notes with a CSV or text export.
Input Multiple local SRT, VTT, or TXT files · Literal search phrase · Optional per-file source linksOutput Matching passages with surrounding context · Cue timestamps when supplied · CSV and TXT search exports
Add a local collection
Up to 50 files, 5 MiB per file, 10 MiB total, 2 million characters per file, and 10,000 combined cues or text lines.
Find the passage
Literal substring matching, within each cue or text line. Phrases can span a subtitle’s wrapped lines, but not separate cues. Optional source links below are supplied by you, never fetched.
Collection and optional source links
No transcripts added yet.
Add transcripts to start searching.
Search results
Matching passages will appear here.
Only 50 matching passages render at once. Each on-screen text excerpt is capped at 2,000 characters; both exports include every match and its full context.
How to use it
- Select several local transcript files or add pasted text to the collection. Each file is validated before it is added, so a broken subtitle block cannot quietly hide a result.
- Optionally enter a source URL beside each transcript. For recognized YouTube video links, timed matches open at the cue’s start second. Other HTTP or HTTPS links remain the source link you supplied.
- Enter a word or phrase, choose case matching and the amount of context, then search. Browse 50 passages at a time or download every match with filenames, timestamps, context, and source links.
A small example
Load the two-file example and search for “question.” The SRT contributes three timestamped passages, and the TXT notes contribute two untimed lines. Search for “a good question” to match a phrase split across wrapped lines inside one subtitle cue.
What to know
- Search is literal substring matching, not semantic search or a regular-expression engine. Phrases do not cross separate subtitle cues or separate TXT lines. Matching is not accent-insensitive.
- Add up to 50 files, at most 5 MiB and 2 million characters each, 10 MiB for the batch, and 10,000 combined subtitle cues or text lines. TXT and subtitle files must be UTF-8.
- Source links are optional and supplied manually. The tool does not look up a video, fetch captions, validate whether a source matches the transcript, or infer timings for TXT.
- The result list shows one passage per matching cue or line, with an occurrence count. Only 50 passages render at once and each text excerpt is capped at 2,000 characters on screen; CSV and TXT exports include every result and full context.
- Subtitle formatting is treated as text and recognized tags are removed for matching. Timing, style, or parse warnings remain visible. Uploaded text is held only in this page’s memory.
References & format documentation
For assistants and developers
This is a browser tool with visible, labeled controls and local processing. Read the usage guide or machine-readable catalog for supported inputs, outputs, examples, and limits. No API key or server endpoint is available. A catalog listing is documentation, not automatic integration with every AI assistant.