You have the words. They came from Zoom, Descript, Otter, or an SRT export, and they are almost ready for show notes or a blog recap. What you do not have is copy that reads like a human edited it. Speaker tags sit in brackets, lines break mid-thought, and capitalization jumps between ALL CAPS and lowercase. Cleaning the transcript once saves rework across every channel that reuses the same script.
What transcript cleaning fixes
Rough transcripts usually carry artifacts from capture and export, not from the speaker. Common issues include:
- Speaker labels such as
[Host]orINTERVIEWER:at the start of lines - Inline timestamps left over from review workflows
- Soft line breaks where the speech-to-text engine wrapped text every few words
- Inconsistent casing when cues were exported from caption tools
- Extra blank lines between short paragraphs
The Case Converter & Transcript Cleaner applies those fixes in one pass. Processing runs in your browser, so files are not uploaded to a remote queue.
Step-by-step workflow
1. Start from the right source file
If your source is an SRT caption file and you only need dialogue, you can paste SRT directly into the cleaner when Extract dialogue from SRT is enabled. The tool drops cue numbers and timing lines.
When you want a plain transcript first, run the SRT to TXT Converter, then paste the TXT into the cleaner for heavier formatting work.
2. Choose capitalization
Pick the mode that matches your deliverable:
| Mode | Best for |
|---|---|
| Sentence case | Show notes, articles, and LMS handouts |
| Title case | Pull quotes used as section headings |
| Keep original casing | Legal names, acronyms, or brand terms that must stay exact |
| lowercase / UPPERCASE | Rare styling needs or fixing an entire broken export |
Sentence case is the default for most publishing workflows because it restores normal reading flow after auto captions.
3. Turn on label and spacing cleanup
For interview and panel content, enable Strip speaker labels and Strip inline timestamps. Enable Unwrap soft line breaks and Normalize spaces when the text looks like a narrow column.
Keep Collapse blank lines on so you retain paragraph breaks without large gaps.
4. Preview before aggressive options
Strip filler words removes standalone tokens such as um and uh. That helps draft polish for narration scripts, but it can change tone. Read the output aloud once before you publish or send to a client.
5. Copy or download
Copy the cleaned text into your CMS, doc, or email. Download .txt when you need a file for batch workflows or to attach to a ticket.
Uploading multiple files
Course teams and podcast producers often clean a batch of episode transcripts with the same settings. Upload several .txt, .srt, or .vtt files, apply one option set, then download individually or as a ZIP archive.
When to use other transcription tools
Cleaning fixes readability. It does not replace the rest of your caption pipeline:
- SRT to TXT when you only need to extract dialogue from subtitles
- TXT to SRT when the script changed and you need new timed cues
- SRT Time Shifter when words are correct but timestamps are early or late
Privacy for draft content
Unreleased lessons, client interviews, and internal recordings often live in transcript files long before public launch. Browser-based cleaning keeps sensitive dialogue on your device while you normalize text for reviewers and stakeholders.
Quick checklist before publish
- Scan the first and last paragraphs for names and product terms after sentence case runs.
- Confirm speaker labels you still need in the narrative were not stripped.
- Spot-check quotes you plan to pull for social posts.
- Archive the rough source file alongside the cleaned export for audit trails.
Clean copy is easier to search, quote, and translate. One polishing pass in the browser is usually faster than fixing the same issues separately in every downstream doc.